Merge branch 'main' into refactor/ai-driver-layout

The framed top-up passes the deployment's dialog the app the framing page's
origin resolves to, as the framed upgrade does. The dashboard link puter.js
opens when no frame can show carries the app in `topup_app`.

.github/workflows/ai-provider-integration-tests.yaml

@@ -31,9 +31,10 @@ jobs:
               run: npm ci
 
             - name: Run AI provider integration tests
-              # Only the provider tests: the other `*.integration.test.ts`
-              # files need no credentials and run with the backend suite.
-              run: npm run test:backend -- $(git ls-files 'src/backend/drivers/ai/**/*.integration.test.ts')
+              # Vitest treats positional args as filename substring filters.
+              # The pattern matches every `*.integration.test.ts` file and
+              # nothing else — the regular backend test suite stays out.
+              run: npm run test:backend -- integration.test
               env:
                   CI: 'true'
                   PUTER_TEST_AI_CLAUDE_API_KEY: ${{ secrets.PUTER_TEST_AI_CLAUDE_API_KEY }}
@@ -48,15 +49,6 @@ jobs:
                   PUTER_TEST_AI_MOONSHOT_API_KEY: ${{ secrets.PUTER_TEST_AI_MOONSHOT_API_KEY }}
                   PUTER_TEST_AI_ZAI_API_KEY: ${{ secrets.PUTER_TEST_AI_ZAI_API_KEY }}
                   PUTER_TEST_AI_ELEVENLABS_API_KEY: ${{ secrets.PUTER_TEST_AI_ELEVENLABS_API_KEY }}
-                  PUTER_TEST_AI_ALIBABA_API_KEY: ${{ secrets.PUTER_TEST_AI_ALIBABA_API_KEY }}
-                  PUTER_TEST_AI_AZURE_OPENAI_API_KEY: ${{ secrets.PUTER_TEST_AI_AZURE_OPENAI_API_KEY }}
-                  PUTER_TEST_AI_AZURE_OPENAI_API_URL: ${{ secrets.PUTER_TEST_AI_AZURE_OPENAI_API_URL }}
-                  PUTER_TEST_AI_BYTEPLUS_API_KEY: ${{ secrets.PUTER_TEST_AI_BYTEPLUS_API_KEY }}
-                  PUTER_TEST_AI_HOONIFY_API_KEY: ${{ secrets.PUTER_TEST_AI_HOONIFY_API_KEY }}
-                  PUTER_TEST_AI_INFRON_API_KEY: ${{ secrets.PUTER_TEST_AI_INFRON_API_KEY }}
-                  PUTER_TEST_AI_KLING_API_KEY: ${{ secrets.PUTER_TEST_AI_KLING_API_KEY }}
-                  PUTER_TEST_AI_META_API_KEY: ${{ secrets.PUTER_TEST_AI_META_API_KEY }}
-                  PUTER_TEST_AI_MINIMAX_API_KEY: ${{ secrets.PUTER_TEST_AI_MINIMAX_API_KEY }}
 
             # Realtime conversations through the SDK against live OpenAI. Builds
             # the SDK bundle these tests load itself, so it runs even when the

config.template.jsonc

@@ -62,6 +62,10 @@
     // direct socket peer. NEVER set to `true` in prod — it trusts every hop
     // and makes X-Forwarded-For forgeable.
     "trust_proxy": false,
+    // Request header a TLS-terminating proxy in front of the server sets to
+    // the client's TLS fingerprint (JA3/JA4). Only set it when that proxy
+    // overwrites any value the client sends. Unset disables it.
+    // "tlsFingerprintHeader": "x-tls-fingerprint",
     "no_browser_launch": false,
 
     // ── Dev watcher (devWatcher extension) ──────────────────────────────

doc/ai-drivers.md

@@ -131,11 +131,10 @@ What the suites assert:
 - A malformed request never fails with an error that has no status, which `/drivers/call` would report as an unclassified 500.
 - For chat, with the vendors answering, a malformed request also never fails with a 5xx and never marks a route unhealthy.
 - When the upstream misbehaves, chat ends with an answer or an `HttpError`. A stream only has to end: the error line a stream carries isn't checked, so a malformed event can still end one with a 500.
-- Requests reach the vendors, and with the vendors answering some chat requests are answered, so a suite that stops exercising the providers fails.
 
 The vendors are played at the HTTP boundary by `upstream` in [shared/fuzzTestUtil.ts](../src/backend/drivers/ai/shared/fuzzTestUtil.ts), which stands in for global `fetch`, undici `fetch` and axios, and makes `https.request` throw so an SDK with its own client fails the suite. The vendors' SDKs run, except Groq's, which is swapped for the OpenAI SDK, and the AWS clients, whose `send` is stubbed. The media suite calls the drivers as a user with unlimited metering, so a request gets past file input and the credit hold to the vendor. A provider that loads its catalog from the vendor registers no models there, so the suites skip it.
 
-A failure lists each distinct finding with its case number and input. Each case is generated from the seed and its own number, so the same case fails again on the next run. Routes are picked by rendezvous hashing: a model added to a catalog takes only the cases it wins, and every other case keeps its route. To search further, raise the case count:
+A failure lists each distinct finding with its case number and input. The random source is seeded, so the same case fails again on the next run. To search further, raise the case count:
 
 ```sh
 PUTER_FUZZ_CASES=100000 npx vitest run --config src/backend/vitest.config.ts src/backend/drivers/ai/interfaces/chat/ChatCompletionDriver.fuzz.test.ts

src/backend/clients/database/SqliteDatabaseClient.test.ts

@@ -27,7 +27,7 @@ import { DatabaseClientFactory } from './index.js';
 import { SqliteDatabaseClient } from './SqliteDatabaseClient.js';
 
 /** Highest schema version the migration table can reach. */
-const CURRENT_SCHEMA_VERSION = 99;
+const CURRENT_SCHEMA_VERSION = 101;
 
 /**
  * These suites migrate real files on disk. Idle they finish in well under a
@@ -735,7 +735,12 @@ describe('SqliteDatabaseClient — boot and migrations', { timeout: DISK_MIGRATI
                     '`recipient_email`) VALUES (?, ?, ?)',
                 ['share-invite', user.id, '[email protected]'],
             );
-            // The full schema as it stood before 0103, so only 0103 reruns.
+            // The full schema as it stood before 0103, so 0103 reruns; 0104
+            // is undone so it can apply again after it.
+            await before.write('DROP INDEX idx_user_signup_tls_fingerprint');
+            await before.write(
+                'ALTER TABLE `user` DROP COLUMN `signup_tls_fingerprint`',
+            );
             await before.write('PRAGMA user_version = 98');
             before.onServerShutdown();
 

src/backend/clients/database/SqliteDatabaseClient.ts

@@ -138,6 +138,8 @@ const AVAILABLE_MIGRATIONS: [number, string[]][] = [
     [96, ['0101_user-scoped-handlers.sql']],
     [97, ['0102_notification-scope-id-index.sql']],
     [98, ['0103_share-entry-owner-backfill.sql']],
+    [99, ['0104_signup-tls-fingerprint.sql']],
+    [100, ['0105_sessions_kind_debug_app.sql']],
 ];
 
 export class SqliteDatabaseClient extends AbstractDatabaseClient {

src/backend/clients/database/migrations/mysql/mysql_mig_58.sql

@@ -0,0 +1,51 @@
+-- Copyright (C) 2024-present Puter Technologies Inc.
+--
+-- This file is part of Puter.
+--
+-- Puter is free software: you can redistribute it and/or modify
+-- it under the terms of the GNU Affero General Public License as published
+-- by the Free Software Foundation, either version 3 of the License, or
+-- (at your option) any later version.
+--
+-- This program is distributed in the hope that it will be useful,
+-- but WITHOUT ANY WARRANTY; without even the implied warranty of
+-- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+-- GNU Affero General Public License for more details.
+--
+-- You should have received a copy of the GNU Affero General Public License
+-- along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+-- `user`.`signup_tls_fingerprint`. See
+-- sqlite/0104_signup-tls-fingerprint.sql for the rationale. No per-file
+-- applied-state tracking, so the column and index are added only when missing
+-- and a replay does nothing.
+
+DROP PROCEDURE IF EXISTS _puter_add_user_signup_tls_fingerprint;
+DELIMITER //
+CREATE PROCEDURE _puter_add_user_signup_tls_fingerprint()
+BEGIN
+  IF NOT EXISTS (
+    SELECT 1 FROM INFORMATION_SCHEMA.COLUMNS
+    WHERE TABLE_SCHEMA = DATABASE()
+      AND TABLE_NAME = 'user'
+      AND COLUMN_NAME = 'signup_tls_fingerprint'
+  ) THEN
+    ALTER TABLE `user`
+      ADD COLUMN `signup_tls_fingerprint` varchar(64) DEFAULT NULL,
+      ALGORITHM=INSTANT;
+  END IF;
+  IF NOT EXISTS (
+    SELECT 1 FROM INFORMATION_SCHEMA.STATISTICS
+    WHERE TABLE_SCHEMA = DATABASE()
+      AND TABLE_NAME = 'user'
+      AND INDEX_NAME = 'idx_user_signup_tls_fingerprint'
+  ) THEN
+    ALTER TABLE `user` ADD INDEX `idx_user_signup_tls_fingerprint`
+      (`signup_tls_fingerprint`), ALGORITHM=INPLACE, LOCK=NONE;
+  END IF;
+END//
+DELIMITER ;
+
+CALL _puter_add_user_signup_tls_fingerprint();
+
+DROP PROCEDURE IF EXISTS _puter_add_user_signup_tls_fingerprint;

src/backend/clients/database/migrations/mysql/mysql_mig_59.sql

@@ -0,0 +1,46 @@
+-- Copyright (C) 2024-present Puter Technologies Inc.
+--
+-- This file is part of Puter.
+--
+-- Puter is free software: you can redistribute it and/or modify
+-- it under the terms of the GNU Affero General Public License as published
+-- by the Free Software Foundation, either version 3 of the License, or
+-- (at your option) any later version.
+--
+-- This program is distributed in the hope that it will be useful,
+-- but WITHOUT ANY WARRANTY; without even the implied warranty of
+-- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+-- GNU Affero General Public License for more details.
+--
+-- You should have received a copy of the GNU Affero General Public License
+-- along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+-- Add 'debug_app' to the `sessions.kind` ENUM: a short-lived app session an
+-- operator creates on a user's behalf. Mirrors SQLite migration 0105.
+--
+-- Idempotent: guarded against COLUMN_TYPE so re-running the directory is a
+-- no-op once 'debug_app' is in the ENUM.
+
+DROP PROCEDURE IF EXISTS _puter_sessions_kind_debug_app;
+DELIMITER //
+CREATE PROCEDURE _puter_sessions_kind_debug_app()
+BEGIN
+  IF NOT EXISTS (
+    SELECT 1 FROM INFORMATION_SCHEMA.COLUMNS
+    WHERE TABLE_SCHEMA = DATABASE()
+      AND TABLE_NAME = 'sessions'
+      AND COLUMN_NAME = 'kind'
+      AND FIND_IN_SET('debug_app', REPLACE(REPLACE(REPLACE(COLUMN_TYPE, 'enum(', ''), ')', ''), '''', '')) > 0
+  ) THEN
+    ALTER TABLE `sessions`
+      MODIFY COLUMN `kind`
+        ENUM('web', 'app', 'access_token', 'asset', 'worker', 'debug_app')
+        NOT NULL DEFAULT 'web',
+      ALGORITHM=INSTANT;
+  END IF;
+END//
+DELIMITER ;
+
+CALL _puter_sessions_kind_debug_app();
+
+DROP PROCEDURE IF EXISTS _puter_sessions_kind_debug_app;

src/backend/clients/database/migrations/postgres/postgres_mig_47.sql

@@ -0,0 +1,22 @@
+-- Copyright (C) 2024-present Puter Technologies Inc.
+--
+-- This file is part of Puter.
+--
+-- Puter is free software: you can redistribute it and/or modify
+-- it under the terms of the GNU Affero General Public License as published
+-- by the Free Software Foundation, either version 3 of the License, or
+-- (at your option) any later version.
+--
+-- This program is distributed in the hope that it will be useful,
+-- but WITHOUT ANY WARRANTY; without even the implied warranty of
+-- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+-- GNU Affero General Public License for more details.
+--
+-- You should have received a copy of the GNU Affero General Public License
+-- along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+-- "user".signup_tls_fingerprint. Mirrors SQLite migration 0104. Idempotent
+-- via IF NOT EXISTS.
+
+ALTER TABLE "user" ADD COLUMN IF NOT EXISTS signup_tls_fingerprint varchar(64);
+CREATE INDEX IF NOT EXISTS idx_user_signup_tls_fingerprint ON "user" (signup_tls_fingerprint);

src/backend/clients/database/migrations/postgres/postgres_mig_48.sql

@@ -0,0 +1,24 @@
+-- Copyright (C) 2024-present Puter Technologies Inc.
+--
+-- This file is part of Puter.
+--
+-- Puter is free software: you can redistribute it and/or modify
+-- it under the terms of the GNU Affero General Public License as published
+-- by the Free Software Foundation, either version 3 of the License, or
+-- (at your option) any later version.
+--
+-- This program is distributed in the hope that it will be useful,
+-- but WITHOUT ANY WARRANTY; without even the implied warranty of
+-- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+-- GNU Affero General Public License for more details.
+--
+-- You should have received a copy of the GNU Affero General Public License
+-- along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+-- Add 'debug_app' to the `sessions.kind` CHECK constraint: a short-lived app
+-- session an operator creates on a user's behalf. Mirrors SQLite migration
+-- 0105. Drop-then-add keeps it idempotent across directory replays.
+
+ALTER TABLE sessions DROP CONSTRAINT IF EXISTS sessions_kind_check;
+ALTER TABLE sessions ADD CONSTRAINT sessions_kind_check
+    CHECK (kind IN ('web', 'app', 'access_token', 'asset', 'worker', 'debug_app'));

src/backend/clients/database/migrations/sqlite/0104_signup-tls-fingerprint.sql

@@ -0,0 +1,22 @@
+-- Copyright (C) 2024-present Puter Technologies Inc.
+--
+-- This file is part of Puter.
+--
+-- Puter is free software: you can redistribute it and/or modify
+-- it under the terms of the GNU Affero General Public License as published
+-- by the Free Software Foundation, either version 3 of the License, or
+-- (at your option) any later version.
+--
+-- This program is distributed in the hope that it will be useful,
+-- but WITHOUT ANY WARRANTY; without even the implied warranty of
+-- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+-- GNU Affero General Public License for more details.
+--
+-- You should have received a copy of the GNU Affero General Public License
+-- along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+-- `user`.`signup_tls_fingerprint`: the TLS fingerprint of the request that
+-- created the account, when a fronting proxy supplies one. Indexed so the
+-- accounts created from one TLS stack can be listed.
+ALTER TABLE `user` ADD COLUMN `signup_tls_fingerprint` varchar(64) DEFAULT NULL;
+CREATE INDEX IF NOT EXISTS idx_user_signup_tls_fingerprint ON `user` (`signup_tls_fingerprint`);

src/backend/clients/database/migrations/sqlite/0105_sessions_kind_debug_app.sql

@@ -0,0 +1,89 @@
+-- Copyright (C) 2024-present Puter Technologies Inc.
+--
+-- This file is part of Puter.
+--
+-- Puter is free software: you can redistribute it and/or modify
+-- it under the terms of the GNU Affero General Public License as published
+-- by the Free Software Foundation, either version 3 of the License, or
+-- (at your option) any later version.
+--
+-- This program is distributed in the hope that it will be useful,
+-- but WITHOUT ANY WARRANTY; without even the implied warranty of
+-- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+-- GNU Affero General Public License for more details.
+--
+-- You should have received a copy of the GNU Affero General Public License
+-- along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+-- Add 'debug_app' to the `kind` CHECK constraint on `sessions`. Mirrors the
+-- MySQL ENUM extension in mysql_mig_59.
+--
+-- SQLite cannot ALTER a CHECK constraint in place, so we follow the same
+-- table-rebuild as migration 0056: create `sessions_new` with the corrected
+-- constraint, copy rows, drop the old table, rename, then recreate every
+-- index that lived on the original.
+
+CREATE TABLE `sessions_new` (
+    "id" INTEGER PRIMARY KEY AUTOINCREMENT,
+    "user_id" INTEGER NOT NULL,
+    "uuid" TEXT NOT NULL,
+    "meta" JSON DEFAULT NULL,
+    "created_at" INTEGER DEFAULT 0,
+    "last_activity" INTEGER DEFAULT 0,
+    "kind" TEXT NOT NULL DEFAULT 'web'
+        CHECK (`kind` IN ('web', 'app', 'access_token', 'asset', 'worker', 'debug_app')),
+    "label" TEXT,
+    "parent_session_id" TEXT,
+    "last_ip" TEXT,
+    "last_user_agent" TEXT,
+    "revoked_at" INTEGER,
+    "expires_at" INTEGER,
+    "app_uid" TEXT,
+    "legacy_token_uid" TEXT,
+    "created_via" TEXT,
+    "auth_id" TEXT,
+    "access_token_uid" TEXT,
+    FOREIGN KEY("user_id") REFERENCES "user" ("id") ON DELETE CASCADE ON UPDATE CASCADE
+);
+
+INSERT INTO `sessions_new` (
+    `id`, `user_id`, `uuid`, `meta`, `created_at`, `last_activity`,
+    `kind`, `label`, `parent_session_id`, `last_ip`, `last_user_agent`,
+    `revoked_at`, `expires_at`, `app_uid`, `legacy_token_uid`,
+    `created_via`, `auth_id`, `access_token_uid`
+)
+SELECT
+    `id`, `user_id`, `uuid`, `meta`, `created_at`, `last_activity`,
+    `kind`, `label`, `parent_session_id`, `last_ip`, `last_user_agent`,
+    `revoked_at`, `expires_at`, `app_uid`, `legacy_token_uid`,
+    `created_via`, `auth_id`, `access_token_uid`
+FROM `sessions`;
+
+DROP TABLE `sessions`;
+
+ALTER TABLE `sessions_new` RENAME TO `sessions`;
+
+CREATE INDEX IF NOT EXISTS `idx_sessions_user_revoked`
+    ON `sessions` (`user_id`, `revoked_at`);
+
+CREATE INDEX IF NOT EXISTS `idx_sessions_parent`
+    ON `sessions` (`parent_session_id`);
+
+CREATE UNIQUE INDEX IF NOT EXISTS `idx_sessions_user_app_active`
+    ON `sessions` (`user_id`, `app_uid`)
+    WHERE `kind` = 'app' AND `revoked_at` IS NULL;
+
+CREATE UNIQUE INDEX IF NOT EXISTS `idx_sessions_legacy_token_active`
+    ON `sessions` (`legacy_token_uid`)
+    WHERE `legacy_token_uid` IS NOT NULL AND `revoked_at` IS NULL;
+
+CREATE INDEX IF NOT EXISTS `idx_sessions_kind_user`
+    ON `sessions` (`kind`, `user_id`);
+
+CREATE INDEX IF NOT EXISTS `idx_sessions_access_token_uid`
+    ON `sessions` (`access_token_uid`)
+    WHERE `access_token_uid` IS NOT NULL;
+
+CREATE UNIQUE INDEX IF NOT EXISTS `idx_sessions_user_worker_active`
+    ON `sessions` (`user_id`, IFNULL(`app_uid`, ''), json_extract(`meta`, '$.worker_name'))
+    WHERE `kind` = 'worker' AND `revoked_at` IS NULL;

src/backend/clients/event/types.ts

@@ -241,6 +241,8 @@ export type EventMap = {
         trail_id?: string;
         /** Device signal forwarded verbatim from the signup request body. */
         fingerprint?: string | null;
+        /** TLS fingerprint of the signup request, when a proxy supplies one. */
+        tls_fingerprint?: string | null;
         /**
          * Set by the abuse harness — require SMS phone verification
          * post-signup.

src/backend/clients/prelude/PreludeClient.test.ts

@@ -132,6 +132,28 @@ describe('PreludeClient', () => {
         });
     });
 
+    it('forwards a JA4 TLS fingerprint and drops any other shape', async () => {
+        fetchMock.mockResolvedValue(okJson({ id: 'v', status: 'success' }));
+        const client = makeClient('sk_test');
+        const ja4 = 't13d1516h2_8daaf6152771_02713d6af862';
+
+        await client.createVerification('+14155550123', {
+            tls_fingerprint: ja4,
+        });
+        await client.createVerification('+14155550123', {
+            ip: '203.0.113.7',
+            // A JA3 hash.
+            tls_fingerprint: 'e7d705a3286e19ea42f587b344ee6865',
+        });
+
+        expect(JSON.parse(fetchMock.mock.calls[0][1].body).signals).toEqual({
+            ja4_fingerprint: ja4,
+        });
+        expect(JSON.parse(fetchMock.mock.calls[1][1].body).signals).toEqual({
+            ip: '203.0.113.7',
+        });
+    });
+
     it('omits the signals object entirely when none are supplied', async () => {
         fetchMock.mockResolvedValue(okJson({ id: 'v', status: 'success' }));
         const client = makeClient('sk_test');

src/backend/clients/prelude/PreludeClient.ts

@@ -25,6 +25,8 @@ const PRELUDE_API_BASE = 'https://api.prelude.dev/v2';
 const REQUEST_TIMEOUT_MS = 8000;
 /** OTP length — matches the 6-box code UI. Prelude allows 4–8. */
 const PRELUDE_CODE_SIZE = 6;
+const JA4_SHAPE =
+    /^[tqd][0-9a-z]{2}[di]\d{4}[0-9a-z]{2}_[0-9a-f]{12}_[0-9a-f]{12}$/;
 /**
  * Channel Prelude prioritizes for delivery. RCS is far cheaper than SMS, so we
  * prefer it by default; Prelude routes to the next reachable channel (SMS) when
@@ -134,6 +136,8 @@ export class PreludeClient extends PuterClient {
             ip?: string;
             device_id?: string;
             user_agent?: string;
+            /** Forwarded only when it is a JA4 value. */
+            tls_fingerprint?: string;
             dispatch_id?: string;
         } = {},
     ): Promise<{
@@ -168,6 +172,8 @@ export class PreludeClient extends PuterClient {
         if (signals.ip) sig.ip = signals.ip;
         if (signals.device_id) sig.device_id = signals.device_id;
         if (signals.user_agent) sig.user_agent = signals.user_agent;
+        if (signals.tls_fingerprint && JA4_SHAPE.test(signals.tls_fingerprint))
+            sig.ja4_fingerprint = signals.tls_fingerprint;
         if (Object.keys(sig).length > 0) body.signals = sig;
         // dispatch_id is a top-level field, not a member of `signals`.
         if (signals.dispatch_id) body.dispatch_id = signals.dispatch_id;

src/backend/controllers/auth/AuthController.test.ts

@@ -722,6 +722,36 @@ describe('AuthController.handleSignup', () => {
         expect(audit.ip_fwd).toBe(forged);
     });
 
+    it('records the request TLS fingerprint on the validate hook and the user row', async () => {
+        const username = `s_${uniq()}`;
+        const tlsFingerprint = 't13d1516h2_8daaf6152771_02713d6af862';
+        let seenTlsFingerprint: unknown;
+
+        await withSignupValidateOverride(
+            (event) => {
+                seenTlsFingerprint = (event as { tls_fingerprint?: unknown })
+                    .tls_fingerprint;
+            },
+            async () => {
+                await controller.handleSignup(
+                    {
+                        ...makeReq({
+                            username,
+                            email: `${username}@test.local`,
+                            password: 'correct-horse-battery',
+                        }),
+                        tlsFingerprint,
+                    },
+                    makeRes(),
+                );
+            },
+        );
+
+        expect(seenTlsFingerprint).toBe(tlsFingerprint);
+        const persisted = await server.stores.user.getByUsername(username);
+        expect(persisted!.signup_tls_fingerprint).toBe(tlsFingerprint);
+    });
+
     it('forces phone verification on every signup when always_require_phone_verification is set', async () => {
         const cfg = (controller as { config: Record<string, unknown> }).config;
         const prev = cfg.always_require_phone_verification;

src/backend/controllers/auth/AuthController.ts

@@ -1510,6 +1510,7 @@ export class AuthController extends PuterController {
                     ip,
                     device_id: req.deviceFingerprint ?? undefined,
                     user_agent: userAgent,
+                    tls_fingerprint: req.tlsFingerprint,
                     dispatch_id: dispatchId,
                 },
             );
@@ -4486,6 +4487,34 @@ export class AuthController extends PuterController {
         res.json({ token: guiToken });
     }
 
+    @Post('/auth/first-party-session', {
+        subdomain: 'api',
+        rateLimit: SESSION_LIMIT,
+        // Hands out a session, so only the GUI's own page script may ask, and
+        // only as the user's sign-in itself.
+        guiOriginOnly: true,
+        requireUserActor: true,
+    })
+    async handleFirstPartySession(req: Request, res: Response): Promise<void> {
+        const { origin } = req.body ?? {};
+        if (typeof origin !== 'string' || !origin) {
+            throw new HttpError(400, 'Missing `origin`', {
+                legacyCode: 'bad_request',
+            });
+        }
+        const token = await this.services.auth.createFirstPartySessionToken(
+            req.actor!,
+            origin,
+            {
+                ip: req.ip || req.socket?.remoteAddress,
+                user_agent: req.headers?.['user-agent'],
+                origin: req.headers?.origin,
+                host: req.headers?.host,
+            },
+        );
+        res.json({ token });
+    }
+
     @Get('/session/sync-cookie', {
         rateLimit: SESSION_LIMIT,
         // Installs the session cookie. Only page script on our own origin
@@ -4510,6 +4539,14 @@ export class AuthController extends PuterController {
             res.status(400).end();
             return;
         }
+        // A session handed to a first-party page never gets a cookie.
+        if (req.actor.session.parentUid) {
+            throw new HttpError(
+                403,
+                'This session cannot install a session cookie',
+                { legacyCode: 'forbidden' },
+            );
+        }
         // As in `/get-gui-token`: the actor is the user's row as auth read it.
         const sessionToken = this.services.auth.createSessionTokenForSession(
             req.actor.user as UserRow,

src/backend/controllers/auth/firstPartySession.http.test.ts

@@ -0,0 +1,282 @@
+/*
+ * Copyright (C) 2024-present Puter Technologies Inc.
+ *
+ * This file is part of Puter.
+ *
+ * Puter is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published
+ * by the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <https://www.gnu.org/licenses/>.
+ */
+
+import { afterAll, beforeAll, describe, expect, it } from 'vitest';
+import type { IConfig } from '../../types.js';
+import { setupPuterTestEnv, type PuterTestEnv } from '../../testUtil.js';
+
+const PAGE = 'https://desk.first-party.test';
+// Seeded with godmode on.
+const DEV_CENTER = 'app-0b37f054-07d4-4627-8765-11bd23e889d4';
+
+/**
+ * One of the deployment's own pages, served apart from the GUI, signs in with a
+ * `web` session of its own under the user's sign-in.
+ */
+describe('first-party page sessions', () => {
+    let env: PuterTestEnv;
+    let userId = 0;
+
+    beforeAll(async () => {
+        env = await setupPuterTestEnv({
+            first_party_session_origins: [`${PAGE}/`],
+        } as IConfig);
+        const row = await env.server.stores.user.getByUsername(
+            env.users.user.username,
+        );
+        userId = row!.id;
+    }, 120_000);
+
+    afterAll(async () => {
+        await env?.shutdown();
+    });
+
+    const call = (
+        method: string,
+        path: string,
+        token: string,
+        body?: unknown,
+        headers: Record<string, string> = {},
+    ) =>
+        fetch(new URL(path, env.apiOrigin), {
+            method,
+            headers: {
+                'content-type': 'application/json',
+                authorization: `Bearer ${token}`,
+                ...headers,
+            },
+            ...(body === undefined ? {} : { body: JSON.stringify(body) }),
+        });
+
+    /** A sign-in of its own, so revoking it leaves the shared one up. */
+    const signIn = async () => {
+        const user = await env.server.stores.user.getById(userId);
+        const { gui_token, session } =
+            await env.server.services.auth.createSessionToken(user!);
+        return { token: gui_token, uuid: session.uuid as string };
+    };
+
+    const mint = async (token: string, origin = PAGE) => {
+        const res = await call('POST', '/auth/first-party-session', token, {
+            origin,
+        });
+        expect(res.status).toBe(200);
+        return ((await res.json()) as { token: string }).token;
+    };
+
+    const actorOf = async (token: string) =>
+        (await env.server.services.auth.authenticate(token)).actor;
+
+    const sessions = async (token: string) => {
+        const res = await call('GET', '/auth/list-sessions', token);
+        expect(res.status).toBe(200);
+        return (await res.json()) as Array<Record<string, unknown>>;
+    };
+
+    it('hands the page a session of its own under the sign-in', async () => {
+        const root = await signIn();
+        const token = await mint(root.token);
+
+        const actor = await actorOf(token);
+        expect(actor?.app ?? null).toBeNull();
+        expect(actor?.accessToken ?? null).toBeNull();
+        expect(actor?.session).toMatchObject({
+            kind: 'web',
+            parentUid: root.uuid,
+        });
+        expect(actor?.session?.uid).not.toBe(root.uuid);
+
+        const row = (await sessions(root.token)).find(
+            (s) => s.uuid === actor!.session!.uid,
+        );
+        expect(row).toMatchObject({
+            kind: 'web',
+            parent_session_id: root.uuid,
+            first_party_origin: PAGE,
+            current: false,
+        });
+    });
+
+    it('reuses the session on a second ask', async () => {
+        const root = await signIn();
+        const first = await actorOf(await mint(root.token));
+        const second = await actorOf(await mint(root.token));
+        expect(second?.session?.uid).toBe(first?.session?.uid);
+    });
+
+    it('acts as the account: launches apps, godmode ones included', async () => {
+        const root = await signIn();
+        const token = await mint(root.token);
+
+        const res = await call('POST', '/auth/get-user-app-token', token, {
+            app_uid: DEV_CENTER,
+        });
+        expect(res.status).toBe(200);
+        expect(await res.json()).toMatchObject({ godmode: true });
+    });
+
+    it('refuses an origin that is not listed', async () => {
+        const root = await signIn();
+        const res = await call(
+            'POST',
+            '/auth/first-party-session',
+            root.token,
+            {
+                origin: 'https://elsewhere.test',
+            },
+        );
+        expect(res.status).toBe(403);
+    });
+
+    it('refuses a page on another origin, the listed page included', async () => {
+        const root = await signIn();
+        for (const origin of [PAGE, 'https://elsewhere.test']) {
+            const res = await call(
+                'POST',
+                '/auth/first-party-session',
+                root.token,
+                { origin: PAGE },
+                { origin },
+            );
+            expect(res.status).toBe(403);
+        }
+    });
+
+    it('refuses app tokens, access tokens and a first-party session', async () => {
+        const root = await signIn();
+        const page = await mint(root.token);
+        const app = (await (
+            await call('POST', '/auth/get-user-app-token', root.token, {
+                origin: 'https://some-app.test',
+            })
+        ).json()) as { token: string };
+        const godmode = (await (
+            await call('POST', '/auth/get-user-app-token', root.token, {
+                app_uid: DEV_CENTER,
+            })
+        ).json()) as { token: string };
+
+        for (const token of [page, app.token, godmode.token]) {
+            const res = await call('POST', '/auth/first-party-session', token, {
+                origin: PAGE,
+            });
+            expect(res.status).toBe(403);
+        }
+    });
+
+    it('never gets a session cookie', async () => {
+        const root = await signIn();
+        const page = await mint(root.token);
+
+        const syncCookie = (token: string) =>
+            fetch(new URL('/session/sync-cookie', env.origin), {
+                headers: { authorization: `Bearer ${token}` },
+            });
+        expect((await syncCookie(root.token)).status).toBe(204);
+
+        const res = await syncCookie(page);
+        expect(res.status).toBe(403);
+        expect(res.headers.get('set-cookie')).toBeNull();
+    });
+
+    it('ends with the sign-in it was handed out under', async () => {
+        const root = await signIn();
+        const page = await mint(root.token);
+        expect(await actorOf(page)).toBeDefined();
+
+        await env.server.services.auth.revokeSession(root.uuid);
+        const result = await env.server.services.auth.authenticate(page);
+        expect(result.actor).toBeUndefined();
+        expect(result.reauth?.reason).toBe('session_revoked');
+    });
+
+    it('takes the tokens minted under it along when its sign-in ends', async () => {
+        const root = await signIn();
+        const page = await mint(root.token);
+        const godmode = (await (
+            await call('POST', '/auth/get-user-app-token', page, {
+                app_uid: DEV_CENTER,
+            })
+        ).json()) as { token: string };
+        const scoped = (await (
+            await call('POST', '/auth/create-access-token', godmode.token, {
+                permissions: ['service:foo:ii:read'],
+                expiresIn: '7d',
+            })
+        ).json()) as { token: string };
+        expect(await actorOf(scoped.token)).toBeDefined();
+
+        await env.server.services.auth.revokeSession(root.uuid);
+        for (const token of [godmode.token, scoped.token]) {
+            expect(
+                (await env.server.services.auth.authenticate(token)).actor,
+            ).toBeUndefined();
+        }
+    });
+
+    it('is revoked on its own without ending the sign-in', async () => {
+        const root = await signIn();
+        const page = await mint(root.token);
+        const pageUuid = (await actorOf(page))!.session!.uid;
+
+        await env.server.services.auth.revokeSession(pageUuid);
+        expect(
+            (await env.server.services.auth.authenticate(page)).actor,
+        ).toBeUndefined();
+        expect(await actorOf(root.token)).toBeDefined();
+    });
+
+    it('keeps the whole current sign-in when revoking all others', async () => {
+        const root = await signIn();
+        const page = await mint(root.token);
+        const other = await signIn();
+        const otherPage = await mint(other.token);
+
+        // Asked from the page: its sign-in stays, and so does the page.
+        await env.server.services.auth.revokeAllSessions(
+            (await actorOf(page))!,
+        );
+        expect(await actorOf(page)).toBeDefined();
+        expect(await actorOf(root.token)).toBeDefined();
+        expect(
+            (await env.server.services.auth.authenticate(other.token)).actor,
+        ).toBeUndefined();
+        expect(
+            (await env.server.services.auth.authenticate(otherPage)).actor,
+        ).toBeUndefined();
+
+        // Asked from the sign-in: the page under it stays too.
+        await env.server.services.auth.revokeAllSessions(
+            (await actorOf(root.token))!,
+        );
+        expect(await actorOf(page)).toBeDefined();
+    });
+
+    it('is not counted as a sign-in', async () => {
+        const root = await signIn();
+        const before = await env.server.stores.session.listSignIns(userId, {
+            limit: 100,
+        });
+        await mint(root.token);
+        const after = await env.server.stores.session.listSignIns(userId, {
+            limit: 100,
+        });
+        expect(after.length).toBe(before.length);
+    });
+});

src/backend/controllers/desktop/DesktopController.js

@@ -1,165 +0,0 @@
-/*
- * Copyright (C) 2024-present Puter Technologies Inc.
- *
- * This file is part of Puter.
- *
- * Puter is free software: you can redistribute it and/or modify
- * it under the terms of the GNU Affero General Public License as published
- * by the Free Software Foundation, either version 3 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU Affero General Public License for more details.
- *
- * You should have received a copy of the GNU Affero General Public License
- * along with this program.  If not, see <https://www.gnu.org/licenses/>.
- */
-
-import { HttpError } from '../../core/http/HttpError.js';
-import { PuterController } from '../types.js';
-
-const ALLOWED_LAYOUTS = ['icons', 'details', 'list'];
-const ALLOWED_SORT_BY = ['name', 'size', 'modified', 'type'];
-const ALLOWED_SORT_ORDER = ['asc', 'desc'];
-
-/**
- * Folder view preference routes: layout and sort_by/sort_order, stored on the
- * folder's fsentries row.
- */
-/**
- * Desktop preference writes — layout, sort order. Each persists on every call,
- * and the GUI fires them on direct user action, so a per-minute ceiling well
- * above human speed is enough to catch a stuck client.
- */
-const PREFERENCE_WRITE_LIMIT = {
-    scope: 'desktop-preference',
-    limit: 120,
-    window: 60_000,
-    key: 'user',
-};
-
-export class DesktopController extends PuterController {
-    registerRoutes(router) {
-        // -- Folder layout -------------------------------------------
-
-        router.post(
-            '/set_layout',
-            {
-                subdomain: 'api',
-                requireUserActor: true,
-                allowFullAccessToken: true,
-                rateLimit: PREFERENCE_WRITE_LIMIT,
-            },
-            async (req, res) => {
-                const { item_uid, item_path, layout } = req.body ?? {};
-                if (!layout || !ALLOWED_LAYOUTS.includes(layout)) {
-                    throw new HttpError(
-                        400,
-                        `\`layout\` must be one of: ${ALLOWED_LAYOUTS.join(', ')}`,
-                        { legacyCode: 'bad_request' },
-                    );
-                }
-                await this.#updateFSEntry(
-                    req.actor,
-                    { item_uid, item_path },
-                    { layout },
-                );
-                res.json({});
-            },
-        );
-
-        // -- Folder sort ---------------------------------------------
-
-        router.post(
-            '/set_sort_by',
-            {
-                subdomain: 'api',
-                requireUserActor: true,
-                allowFullAccessToken: true,
-                rateLimit: PREFERENCE_WRITE_LIMIT,
-            },
-            async (req, res) => {
-                const { item_uid, item_path, sort_by, sort_order } =
-                    req.body ?? {};
-                if (!sort_by || !ALLOWED_SORT_BY.includes(sort_by)) {
-                    throw new HttpError(
-                        400,
-                        `\`sort_by\` must be one of: ${ALLOWED_SORT_BY.join(', ')}`,
-                        { legacyCode: 'bad_request' },
-                    );
-                }
-                const resolvedOrder = sort_order ?? 'asc';
-                if (!ALLOWED_SORT_ORDER.includes(resolvedOrder)) {
-                    throw new HttpError(
-                        400,
-                        `\`sort_order\` must be one of: ${ALLOWED_SORT_ORDER.join(', ')}`,
-                        { legacyCode: 'bad_request' },
-                    );
-                }
-                await this.#updateFSEntry(
-                    req.actor,
-                    { item_uid, item_path },
-                    {
-                        sort_by,
-                        sort_order: resolvedOrder,
-                    },
-                );
-                res.json({});
-            },
-        );
-    }
-
-    // -- Helpers ------------------------------------------------------
-
-    /**
-     * Update columns on an actor-owned fsentry. Accepts either `item_uid` (fast
-     * path) or `item_path` (path lookups in FSEntryStore now have a
-     * recursive-CTE fallback for legacy rows with a NULL `path` column, so this
-     * works for old accounts too).
-     *
-     * Ownership: `entry.user_id === actor.user.id`. Kept as added validation
-     * against bad paths — the previous "drop user_id entirely" theory turned
-     * out to be wrong (the actual legacy issue was NULL paths, not user_id
-     * drift), so this filter doesn't lock out old accounts in practice.
-     */
-    async #updateFSEntry(actor, { item_uid, item_path }, patch) {
-        if (!item_uid && !item_path) {
-            throw new HttpError(400, 'Missing `item_uid` or `item_path`', {
-                legacyCode: 'bad_request',
-            });
-        }
-
-        const entry = item_uid
-            ? await this.stores.fsEntry.getEntryByUuid(item_uid)
-            : await this.stores.fsEntry.getEntryByPath(item_path);
-        if (!entry) {
-            throw new HttpError(404, 'Item not found', {
-                legacyCode: 'not_found',
-            });
-        }
-
-        const actorUserId = actor?.user?.id;
-        if (typeof actorUserId !== 'number' || entry.userId !== actorUserId) {
-            throw new HttpError(403, 'Not allowed to update this item', {
-                legacyCode: 'forbidden',
-            });
-        }
-
-        const keys = Object.keys(patch);
-        const setClause = keys.map((k) => `\`${k}\` = ?`).join(', ');
-        const values = keys.map((k) => patch[k]);
-
-        await this.clients.db.write(
-            `UPDATE \`fsentries\` SET ${setClause} WHERE \`id\` = ?`,
-            [...values, entry.id],
-        );
-
-        await this.stores.fsEntry.invalidateEntryCacheByUuid(entry.uuid);
-    }
-
-    onServerStart() {}
-    onServerPrepareShutdown() {}
-    onServerShutdown() {}
-}

src/backend/controllers/desktop/DesktopController.test.ts

@@ -1,381 +0,0 @@
-/*
- * Copyright (C) 2024-present Puter Technologies Inc.
- *
- * This file is part of Puter.
- *
- * Puter is free software: you can redistribute it and/or modify
- * it under the terms of the GNU Affero General Public License as published
- * by the Free Software Foundation, either version 3 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU Affero General Public License for more details.
- *
- * You should have received a copy of the GNU Affero General Public License
- * along with this program.  If not, see <https://www.gnu.org/licenses/>.
- */
-
-import type { Request, RequestHandler, Response } from 'express';
-import { afterAll, beforeAll, describe, expect, it, vi } from 'vitest';
-import { v4 as uuidv4 } from 'uuid';
-import type { Actor } from '../../core/actor.js';
-import { PuterRouter } from '../../core/http/PuterRouter.js';
-import { PuterServer } from '../../server.js';
-import { setupTestServer } from '../../testUtil.js';
-import { generateDefaultFsentries } from '../../util/userProvisioning.js';
-
-// ── Test harness ────────────────────────────────────────────────────
-//
-// Boots one real PuterServer (in-memory sqlite + dynamo + s3 + mock
-// redis) and re-registers DesktopController's inline lambda routes
-// onto a fresh PuterRouter so each handler is reachable. Each test
-// makes its own user via `makeUser` and exercises the live controller
-// against the real wired stores (user, fsEntry) and DB client.
-
-let server: PuterServer;
-let router: PuterRouter;
-
-beforeAll(async () => {
-    server = await setupTestServer();
-    router = new PuterRouter();
-    server.controllers.desktop.registerRoutes(router);
-});
-
-afterAll(async () => {
-    await server?.shutdown();
-});
-
-const makeUser = async (): Promise<{ actor: Actor; userId: number }> => {
-    const username = `dc-${Math.random().toString(36).slice(2, 10)}`;
-    const created = await server.stores.user.create({
-        username,
-        uuid: uuidv4(),
-        password: null,
-        email: `${username}@test.local`,
-        free_storage: 100 * 1024 * 1024,
-        requires_email_confirmation: false,
-    });
-    await generateDefaultFsentries(
-        server.clients.db,
-        server.stores.user,
-        created,
-    );
-    const refreshed = (await server.stores.user.getById(created.id))!;
-    return {
-        userId: refreshed.id,
-        actor: {
-            user: {
-                id: refreshed.id,
-                uuid: refreshed.uuid,
-                username: refreshed.username,
-                email: refreshed.email ?? null,
-                email_confirmed: true,
-            } as Actor['user'],
-        },
-    };
-};
-
-interface CapturedResponse {
-    statusCode: number;
-    body: unknown;
-}
-
-const makeReq = (init: {
-    body?: unknown;
-    actor?: Actor;
-}): Request => {
-    return {
-        body: init.body ?? {},
-        query: {},
-        headers: {},
-        actor: init.actor,
-    } as unknown as Request;
-};
-
-const makeRes = () => {
-    const captured: CapturedResponse = { statusCode: 200, body: undefined };
-    const res = {
-        json: vi.fn((value: unknown) => {
-            captured.body = value;
-            return res;
-        }),
-        status: vi.fn((code: number) => {
-            captured.statusCode = code;
-            return res;
-        }),
-        setHeader: vi.fn(() => res),
-    };
-    return { res: res as unknown as Response, captured };
-};
-
-const findHandler = (method: string, path: string): RequestHandler => {
-    const route = router.routes.find(
-        (r) => r.method === method && r.path === path,
-    );
-    if (!route) throw new Error(`No ${method.toUpperCase()} ${path} route`);
-    return route.handler;
-};
-
-const callRoute = async (
-    method: string,
-    path: string,
-    req: Request,
-    res: Response,
-) => {
-    const handler = findHandler(method, path);
-    await handler(req, res, () => {
-        throw new Error('handler called next() unexpectedly');
-    });
-};
-
-describe('DesktopController routes', () => {
-    it('has no desktop background route', () => {
-        expect(
-            router.routes.some((r) => r.path === '/set-desktop-bg'),
-        ).toBe(false);
-    });
-
-    it('has no taskbar items route', () => {
-        expect(
-            router.routes.some((r) => r.path === '/update-taskbar-items'),
-        ).toBe(false);
-    });
-});
-
-// ── /set_layout ─────────────────────────────────────────────────────
-
-describe('DesktopController POST /set_layout', () => {
-    it('persists the new layout on the matching fsentry', async () => {
-        const { actor } = await makeUser();
-        const username = actor.user!.username!;
-        const documents = await server.stores.fsEntry.getEntryByPath(
-            `/${username}/Documents`,
-        );
-        expect(documents).not.toBeNull();
-
-        const { res, captured } = makeRes();
-        await callRoute(
-            'post',
-            '/set_layout',
-            makeReq({
-                body: { item_uid: documents!.uuid, layout: 'icons' },
-                actor,
-            }),
-            res,
-        );
-        expect(captured.body).toEqual({});
-
-        const refreshed = await server.stores.fsEntry.getEntryByUuid(
-            documents!.uuid,
-        );
-        expect(refreshed?.layout).toBe('icons');
-    });
-
-    it('resolves by item_path when item_uid is omitted', async () => {
-        const { actor } = await makeUser();
-        const username = actor.user!.username!;
-        const path = `/${username}/Pictures`;
-
-        const { res } = makeRes();
-        await callRoute(
-            'post',
-            '/set_layout',
-            makeReq({
-                body: { item_path: path, layout: 'list' },
-                actor,
-            }),
-            res,
-        );
-
-        const refreshed = await server.stores.fsEntry.getEntryByPath(path);
-        expect(refreshed?.layout).toBe('list');
-    });
-
-    it('throws 400 for an unknown layout value', async () => {
-        const { actor } = await makeUser();
-        const username = actor.user!.username!;
-        const documents = await server.stores.fsEntry.getEntryByPath(
-            `/${username}/Documents`,
-        );
-
-        const { res } = makeRes();
-        await expect(
-            callRoute(
-                'post',
-                '/set_layout',
-                makeReq({
-                    body: { item_uid: documents!.uuid, layout: 'gallery' },
-                    actor,
-                }),
-                res,
-            ),
-        ).rejects.toMatchObject({ statusCode: 400 });
-    });
-
-    it('throws 400 when neither item_uid nor item_path is supplied', async () => {
-        const { actor } = await makeUser();
-        const { res } = makeRes();
-        await expect(
-            callRoute(
-                'post',
-                '/set_layout',
-                makeReq({ body: { layout: 'icons' }, actor }),
-                res,
-            ),
-        ).rejects.toMatchObject({ statusCode: 400 });
-    });
-
-    it('throws 404 when the fsentry cannot be resolved', async () => {
-        const { actor } = await makeUser();
-        const { res } = makeRes();
-        await expect(
-            callRoute(
-                'post',
-                '/set_layout',
-                makeReq({
-                    body: {
-                        item_uid: '00000000-0000-0000-0000-000000000000',
-                        layout: 'icons',
-                    },
-                    actor,
-                }),
-                res,
-            ),
-        ).rejects.toMatchObject({ statusCode: 404 });
-    });
-
-    it("throws 403 when the fsentry belongs to another user", async () => {
-        const owner = await makeUser();
-        const intruder = await makeUser();
-        const ownerUsername = owner.actor.user!.username!;
-        const ownerEntry = await server.stores.fsEntry.getEntryByPath(
-            `/${ownerUsername}/Documents`,
-        );
-
-        const { res } = makeRes();
-        await expect(
-            callRoute(
-                'post',
-                '/set_layout',
-                makeReq({
-                    body: { item_uid: ownerEntry!.uuid, layout: 'icons' },
-                    actor: intruder.actor,
-                }),
-                res,
-            ),
-        ).rejects.toMatchObject({ statusCode: 403 });
-
-        // Owner's entry remains untouched.
-        const refreshed = await server.stores.fsEntry.getEntryByUuid(
-            ownerEntry!.uuid,
-        );
-        expect(refreshed?.layout).toBeFalsy();
-    });
-});
-
-// ── /set_sort_by ────────────────────────────────────────────────────
-
-describe('DesktopController POST /set_sort_by', () => {
-    it('persists sort_by + sort_order on the matching fsentry', async () => {
-        const { actor } = await makeUser();
-        const username = actor.user!.username!;
-        const documents = await server.stores.fsEntry.getEntryByPath(
-            `/${username}/Documents`,
-        );
-
-        const { res, captured } = makeRes();
-        await callRoute(
-            'post',
-            '/set_sort_by',
-            makeReq({
-                body: {
-                    item_uid: documents!.uuid,
-                    sort_by: 'name',
-                    sort_order: 'desc',
-                },
-                actor,
-            }),
-            res,
-        );
-        expect(captured.body).toEqual({});
-
-        const refreshed = await server.stores.fsEntry.getEntryByUuid(
-            documents!.uuid,
-        );
-        expect(refreshed?.sortBy).toBe('name');
-        expect(refreshed?.sortOrder).toBe('desc');
-    });
-
-    it('defaults sort_order to "asc" when omitted', async () => {
-        const { actor } = await makeUser();
-        const username = actor.user!.username!;
-        const pictures = await server.stores.fsEntry.getEntryByPath(
-            `/${username}/Pictures`,
-        );
-
-        const { res } = makeRes();
-        await callRoute(
-            'post',
-            '/set_sort_by',
-            makeReq({
-                body: { item_uid: pictures!.uuid, sort_by: 'modified' },
-                actor,
-            }),
-            res,
-        );
-
-        const refreshed = await server.stores.fsEntry.getEntryByUuid(
-            pictures!.uuid,
-        );
-        expect(refreshed?.sortBy).toBe('modified');
-        expect(refreshed?.sortOrder).toBe('asc');
-    });
-
-    it('throws 400 for an unknown sort_by value', async () => {
-        const { actor } = await makeUser();
-        const username = actor.user!.username!;
-        const docs = await server.stores.fsEntry.getEntryByPath(
-            `/${username}/Documents`,
-        );
-
-        const { res } = makeRes();
-        await expect(
-            callRoute(
-                'post',
-                '/set_sort_by',
-                makeReq({
-                    body: { item_uid: docs!.uuid, sort_by: 'random' },
-                    actor,
-                }),
-                res,
-            ),
-        ).rejects.toMatchObject({ statusCode: 400 });
-    });
-
-    it('throws 400 for an invalid sort_order', async () => {
-        const { actor } = await makeUser();
-        const username = actor.user!.username!;
-        const docs = await server.stores.fsEntry.getEntryByPath(
-            `/${username}/Documents`,
-        );
-
-        const { res } = makeRes();
-        await expect(
-            callRoute(
-                'post',
-                '/set_sort_by',
-                makeReq({
-                    body: {
-                        item_uid: docs!.uuid,
-                        sort_by: 'name',
-                        sort_order: 'sideways',
-                    },
-                    actor,
-                }),
-                res,
-            ),
-        ).rejects.toMatchObject({ statusCode: 400 });
-    });
-});

src/backend/controllers/drivers/DriverController.concurrent.test.ts

@@ -324,7 +324,7 @@ describe('DriverController — concurrent acquire/release', () => {
         const driver = {
             ...makeSyntheticDriver(),
             ping: async () => {
-                Context.get('keepConcurrencySlot')?.(
+                Context.get('keepDriverSlot')?.(
                     new Promise<void>((resolve) => {
                         settleWork = resolve;
                     }),

src/backend/controllers/drivers/DriverController.ts

@@ -51,7 +51,7 @@ import { PermissionUtil } from '../../services/permission/permissionUtil.js';
 import type { WithLifecycle } from '../../types';
 import { withSpan } from '../../util/span.js';
 import { PuterController } from '../types.js';
-import { CallSlot } from '../../core/http/callSlot.js';
+import { CallSlot } from './callSlot.js';
 
 type DriverInstance = WithLifecycle & Record<string, unknown>;
 
@@ -446,7 +446,7 @@ export class DriverController extends PuterController {
                 });
             }
             slot.attach(handle, res);
-            Context.set('keepConcurrencySlot', (work) => slot.keep(work));
+            Context.set('keepDriverSlot', (work) => slot.keep(work));
         }
 
         // Stash the requested driver name in Context so multi-provider

src/backend/controllers/drivers/callSlot.ts

@@ -18,6 +18,7 @@
  */
 
 import type { Response } from 'express';
+import { onResponseClose } from '../../core/http/onResponseClose.js';
 
 // Under the concurrency store's one-hour orphan TTL.
 const SLOT_RENEW_MS = 10 * 60 * 1000;
@@ -28,9 +29,9 @@ interface SlotHandle {
 }
 
 /**
- * A request's concurrency slot, held until the response is done and the work
- * registered with `keep` has settled, including work that carries on after the
- * caller hung up or was answered.
+ * A driver call's concurrency slot, held until the response is done and the
+ * call's work has settled, including work a driver carries on after the caller
+ * hung up or was answered.
  */
 export class CallSlot {
     #handle?: SlotHandle;
@@ -46,7 +47,7 @@ export class CallSlot {
             this.#maybeRelease();
         };
         res.once('finish', responded);
-        res.once('close', responded);
+        onResponseClose(res, responded);
         if (res.destroyed || res.writableFinished) responded();
         if (!this.#released && this.#work.size > 0) this.#startRenewing();
     }

src/backend/controllers/index.ts

@@ -21,7 +21,6 @@ import { AppController } from './apps/AppController.js';
 import { AppFeedbackController } from './feedback/AppFeedbackController.js';
 import { AuthController } from './auth/AuthController.js';
 import { BroadcastController } from './broadcast/BroadcastController.js';
-import { DesktopController } from './desktop/DesktopController.js';
 import { DriverController } from './drivers/DriverController.js';
 import { EventsController } from './events/EventsController.js';
 import { FSController } from './fs/FSController.js';
@@ -49,7 +48,6 @@ export const puterControllers = {
     profile: ProfileController,
     apps: AppController,
     appFeedback: AppFeedbackController,
-    desktop: DesktopController,
     hosting: HostingController,
     system: SystemController,
     fs: FSController,

src/backend/controllers/puterai/PuterAIController.ts

@@ -1304,18 +1304,13 @@ export class PuterAIController extends PuterController {
         return driver;
     }
 
-    /**
-     * A completion whose caller may hang up. It runs on to its usage report
-     * either way, so it holds the route's concurrency slot until it settles.
-     */
+    /** A completion that stops, and stops billing, when the caller hangs up. */
     #complete(
         res: Response,
         args: ICompleteArguments,
     ): Promise<IChatCompleteResult> {
         Context.set('abortSignal', abortOnDisconnect(res));
-        const completion = this.#driver().complete(args);
-        Context.get('keepConcurrencySlot')?.(completion);
-        return completion;
+        return this.#driver().complete(args);
     }
 }
 

src/backend/controllers/puterai/sse.ts

@@ -26,6 +26,7 @@
  */
 
 import type { Response } from 'express';
+import { onResponseClose } from '../../core/http/onResponseClose.js';
 import type { UsageDetails } from '../../drivers/ai/interfaces/chat/types.js';
 import { promoteStopForToolCalls } from '../../drivers/ai/interfaces/chat/utils/stopReason.js';
 import { anthropicUsage } from './anthropicWire.js';
@@ -79,7 +80,7 @@ export const startSse = (res: Response, opts: { ping?: boolean } = {}): Sse => {
         timer.unref?.();
     }
 
-    res.once('close', () => {
+    onResponseClose(res, () => {
         ended = true;
         if (timer) clearInterval(timer);
     });

src/backend/controllers/webdav/WebDAVController.get.http.test.ts

@@ -0,0 +1,210 @@
+/*
+ * Copyright (C) 2024-present Puter Technologies Inc.
+ *
+ * This file is part of Puter.
+ *
+ * Puter is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published
+ * by the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <https://www.gnu.org/licenses/>.
+ */
+
+import { GetObjectCommand } from '@aws-sdk/client-s3';
+import { randomBytes } from 'node:crypto';
+import { Readable } from 'node:stream';
+import { afterAll, beforeAll, describe, expect, it, vi } from 'vitest';
+import {
+    RESPONSE_CLOSE_LISTENER_BUDGET,
+    setupPuterTestEnv,
+    TEST_USER_CREDENTIALS,
+    trackResponseCloseListeners,
+    type PuterTestEnv,
+} from '../../testUtil.js';
+
+/** File downloads over WebDAV: the bytes, ranges, and either end failing. */
+describe('WebDAV GET over HTTP', () => {
+    let env: PuterTestEnv;
+
+    beforeAll(async () => {
+        env = await setupPuterTestEnv();
+    }, 60_000);
+
+    afterAll(async () => {
+        await env?.shutdown();
+    });
+
+    const dav = (path: string, init: RequestInit = {}) => {
+        const port = new URL(env.apiOrigin).port;
+        const auth = Buffer.from(
+            `${TEST_USER_CREDENTIALS.username}:${TEST_USER_CREDENTIALS.password}`,
+        ).toString('base64');
+        return fetch(`http://dav.puter.localhost:${port}${encodeURI(path)}`, {
+            ...init,
+            headers: {
+                Authorization: `Basic ${auth}`,
+                // Uncompressed, so the bytes on the wire are the stored ones.
+                'Accept-Encoding': 'identity',
+                ...(init.headers as Record<string, string> | undefined),
+            },
+        });
+    };
+
+    const writeFile = async (name: string, content: Buffer) => {
+        const { username } = env.users.user;
+        const user = await env.server.stores.user.getByUsername(username);
+        const path = `/${username}/Documents/${name}`;
+        const { fsEntry } = await env.server.services.fs.write(user!.id, {
+            fileMetadata: {
+                path,
+                size: content.byteLength,
+                contentType: 'application/octet-stream',
+            },
+            fileContent: content,
+        });
+        return { path, uuid: fsEntry.uuid };
+    };
+
+    it('returns the stored bytes, whole or as a range', async () => {
+        const content = randomBytes(128 * 1024 + 3);
+        const file = await writeFile('dav-get.bin', content);
+
+        const whole = await dav(file.path);
+        expect(whole.status).toBe(200);
+        expect(whole.headers.get('content-length')).toBe(
+            String(content.byteLength),
+        );
+        expect(Buffer.from(await whole.arrayBuffer())).toEqual(content);
+
+        const range = await dav(file.path, {
+            headers: { Range: 'bytes=10-19' },
+        });
+        expect(range.status).toBe(206);
+        expect(range.headers.get('content-range')).toBe(
+            `bytes 10-19/${content.byteLength}`,
+        );
+        expect(Buffer.from(await range.arrayBuffer())).toEqual(
+            content.subarray(10, 20),
+        );
+    });
+
+    /**
+     * Storage hands over one chunk of `file` and then goes quiet, as a large
+     * object does while the client is still reading.
+     */
+    const stallStorage = (file: { uuid: string }) => {
+        const s3 = env.server.stores.s3Object;
+        const client = env.server.clients.s3.get(s3.resolveRegion(null));
+        const send = client.send.bind(client);
+        const bodies: Readable[] = [];
+        const spy = vi.spyOn(client, 'send').mockImplementation(((
+            command: unknown,
+        ) => {
+            if (
+                !(command instanceof GetObjectCommand) ||
+                command.input.Key !== file.uuid
+            ) {
+                return send(command as never);
+            }
+            let pushed = false;
+            const body = new Readable({
+                read() {
+                    if (!pushed) {
+                        pushed = true;
+                        this.push(Buffer.alloc(16 * 1024, 1));
+                    }
+                },
+            });
+            bodies.push(body);
+            return Promise.resolve({
+                Body: body,
+                ContentLength: 1024 * 1024,
+                ContentType: 'application/octet-stream',
+            });
+        }) as never);
+        return { bodies, restore: () => spy.mockRestore() };
+    };
+
+    const expectFullDownload = async (path: string) => {
+        const response = await dav(path);
+        expect(response.status).toBe(200);
+        expect((await response.arrayBuffer()).byteLength).toBe(1024 * 1024);
+    };
+
+    it('releases the storage stream when the client aborts', async () => {
+        const file = await writeFile(
+            'dav-abort.bin',
+            Buffer.alloc(1024 * 1024, 1),
+        );
+        const storage = stallStorage(file);
+        try {
+            const abort = new AbortController();
+            const response = await dav(file.path, { signal: abort.signal });
+            expect(response.status).toBe(200);
+            await response.body!.getReader().read();
+            abort.abort();
+
+            expect(storage.bodies).toHaveLength(1);
+            await vi.waitFor(() =>
+                expect(storage.bodies[0]!.destroyed).toBe(true),
+            );
+        } finally {
+            storage.restore();
+        }
+
+        await expectFullDownload(file.path);
+    });
+
+    it('cuts the download off when storage fails mid-stream', async () => {
+        const file = await writeFile(
+            'dav-reset.bin',
+            Buffer.alloc(1024 * 1024, 1),
+        );
+        const storage = stallStorage(file);
+        try {
+            const response = await dav(file.path);
+            expect(response.status).toBe(200);
+            const reader = response.body!.getReader();
+            await reader.read();
+            storage.bodies[0]!.destroy(new Error('upstream reset'));
+
+            const outcome = await (async () => {
+                try {
+                    while (!(await reader.read()).done);
+                    return 'ended';
+                } catch {
+                    return 'cut off';
+                }
+            })();
+            expect(outcome).toBe('cut off');
+        } finally {
+            storage.restore();
+        }
+
+        await expectFullDownload(file.path);
+    });
+
+    it('stays within the close listener budget per response', async () => {
+        const file = await writeFile('dav-listeners.bin', randomBytes(4096));
+        const tracker = trackResponseCloseListeners();
+        try {
+            const response = await dav(file.path);
+            expect(response.status).toBe(200);
+            await response.arrayBuffer();
+        } finally {
+            tracker.restore();
+        }
+        const peak = tracker.peak(
+            (req) => req.method === 'GET' && req.url === file.path,
+        );
+        expect(peak).toBeGreaterThan(0);
+        expect(peak).toBeLessThanOrEqual(RESPONSE_CLOSE_LISTENER_BUDGET);
+    });
+});

src/backend/controllers/webdav/WebDAVController.test.ts

@@ -1365,6 +1365,7 @@ describe('WebDAVController verbs', () => {
                 callback();
             },
         });
+        sink.on('error', () => {});
         sink.on('finish', () => {
             captured.ended = true;
         });

src/backend/controllers/webdav/WebDAVController.ts

@@ -21,7 +21,6 @@ import type { Team2faLookup } from '../../core/http/middleware/gates.js';
 import { compare as bcryptCompare } from 'bcrypt';
 import type { Request, Response } from 'express';
 import { posix as pathPosix } from 'node:path';
-import { pipeline } from 'node:stream/promises';
 import type { Actor } from '../../core/actor.js';
 import { HttpError } from '../../core/http/HttpError.js';
 import { pipeToResponse } from '../../core/http/pipeToResponse.js';
@@ -551,12 +550,7 @@ export class WebDAVController extends PuterController {
                 'Content-Length': String(result.contentLength ?? 0),
             });
         }
-        try {
-            await pipeline(result.body, res);
-        } catch {
-            // Client disconnect or upstream stream error: pipeline already
-            // tore down both ends, so nothing is left dangling to crash on.
-        }
+        await pipeToResponse(result.body, res);
     }
 
     // -- PROPFIND ----------------------------------------------------

src/backend/core/actor.ts

@@ -76,8 +76,14 @@ export interface Actor {
     /**
      * Set when authenticated by a session token or an app-under-user token
      * carrying one. `kind` mirrors the session row (`web`, `app`, `worker`).
+     * `parentUid` is the sign-in a `web` session was handed out under, set only
+     * on a first-party page's session.
      */
-    session?: { uid: string; kind?: string | null } | null;
+    session?: {
+        uid: string;
+        kind?: string | null;
+        parentUid?: string | null;
+    } | null;
     /**
      * Events handler runs behind this actor's writes; set only from an events
      * handler token or a token one minted. Provenance, not authority, so

src/backend/core/context.ts

@@ -69,11 +69,11 @@ export interface KnownContextFields {
      */
     abortSignal: AbortSignal;
     /**
-     * Keeps the request's concurrency slot until `work` settles, for work that
-     * outlives the response. Set by DriverController for a capped method and by
-     * the route concurrency gate.
+     * Keeps the call's concurrency slot until `work` settles (set by
+     * DriverController when the method is capped), for work that outlives the
+     * response.
      */
-    keepConcurrencySlot: (work: Promise<unknown>) => void;
+    keepDriverSlot: (work: Promise<unknown>) => void;
     /**
      * Aborts the chat driver's upstream request. A client disconnect doesn't
      * fire it: the upstream bills a request it has accepted either way, so the
@@ -102,7 +102,7 @@ const KNOWN_KEY_FLAGS: Record<keyof KnownContextFields, true> = {
     requestId: true,
     driverName: true,
     abortSignal: true,
-    keepConcurrencySlot: true,
+    keepDriverSlot: true,
     upstreamSignal: true,
     strictUpstreamErrors: true,
     platformWrite: true,

src/backend/core/http/abortOnDisconnect.ts

@@ -19,6 +19,7 @@
 
 import type { Response } from 'express';
 import { HttpError } from './HttpError.js';
+import { onResponseClose } from './onResponseClose.js';
 
 /**
  * A signal that fires when the caller hangs up before its response is done, so
@@ -38,6 +39,6 @@ export const abortOnDisconnect = (res: Response): AbortSignal => {
         }
     };
     if (res.destroyed) onClose();
-    else res.once('close', onClose);
+    else onResponseClose(res, onClose);
     return abort.signal;
 };

src/backend/core/http/expressAugmentation.ts

@@ -104,6 +104,13 @@ declare global {
              */
             deviceFingerprint?: string;
 
+            /**
+             * The client's TLS fingerprint, read from the header named by
+             * `tlsFingerprintHeader`. Shared by every client on the same TLS
+             * stack, so it identifies software, not a person.
+             */
+            tlsFingerprint?: string;
+
             /**
              * Parsed cookies, populated by the global `cookie-parser`
              * middleware.

src/backend/core/http/middleware/egressMetering.ts

@@ -28,6 +28,7 @@ import {
 } from '../../../services/metering/costs.js';
 import type { UsageInput } from '../../../services/metering/types';
 import '../expressAugmentation';
+import { onResponseClose } from '../onResponseClose';
 
 /**
  * Subset of the metering service this middleware needs. Metering is optional
@@ -195,7 +196,7 @@ export const createEgressMeteringMiddleware = (
         // 'close' rather than 'finish': it fires for a response that completed
         // AND for one whose connection died mid-stream, and an aborted download
         // still sent whatever it sent.
-        res.once('close', () => {
+        onResponseClose(res, () => {
             try {
                 if (!isMeteredHost(req)) return;
                 const actor = resolveEgressActor(req);

src/backend/core/http/middleware/fingerprint.test.ts

@@ -23,8 +23,11 @@ import { createFingerprintMiddleware } from './fingerprint';
 
 const VALID_FP = 'dev-fingerprint-abc123';
 
-const run = (reqPartial: Partial<Request> & Record<string, unknown>) => {
-    const middleware = createFingerprintMiddleware();
+const run = (
+    reqPartial: Partial<Request> & Record<string, unknown>,
+    opts?: Parameters<typeof createFingerprintMiddleware>[0],
+) => {
+    const middleware = createFingerprintMiddleware(opts);
     const req = {
         headers: {},
         ...reqPartial,
@@ -90,4 +93,21 @@ describe('createFingerprintMiddleware', () => {
         const { req } = run({ ip: '203.0.113.1' });
         expect(req.deviceFingerprint).toBeUndefined();
     });
+
+    it('reads the TLS fingerprint from the configured header only', () => {
+        const ja4 = 't13d1516h2_8daaf6152771_02713d6af862';
+        const headers = { 'x-tls-fp': ja4 };
+
+        expect(run({ headers }).req.tlsFingerprint).toBeUndefined();
+        expect(
+            run({ headers }, { tlsFingerprintHeader: 'X-TLS-FP' }).req
+                .tlsFingerprint,
+        ).toBe(ja4);
+        expect(
+            run(
+                { headers: { 'x-tls-fp': 'not a fingerprint!' } },
+                { tlsFingerprintHeader: 'x-tls-fp' },
+            ).req.tlsFingerprint,
+        ).toBeUndefined();
+    });
 });

src/backend/core/http/middleware/fingerprint.ts

@@ -33,6 +33,20 @@ const DEVICE_FINGERPRINT_SHAPE = /^[A-Za-z0-9._-]{8,128}$/;
 /** Header the GUI may send the device fingerprint on for non-signup requests. */
 const DEVICE_FINGERPRINT_HEADER = 'x-puter-device-fingerprint';
 
+/** JA3 / JA4-style values; anything else is dropped. */
+const TLS_FINGERPRINT_SHAPE = /^[A-Za-z0-9_]{10,64}$/;
+
+function readTlsFingerprint(
+    req: { headers?: Record<string, unknown> },
+    header: string | undefined,
+): string | undefined {
+    if (!header) return undefined;
+    const value = req.headers?.[header];
+    return typeof value === 'string' && TLS_FINGERPRINT_SHAPE.test(value)
+        ? value
+        : undefined;
+}
+
 function readDeviceFingerprint(req: {
     body?: unknown;
     headers?: Record<string, unknown>;
@@ -65,15 +79,22 @@ function readDeviceFingerprint(req: {
  *   and spoofable, but stable per real device across IP rotation. The rate
  *   limiter's 'fingerprint' strategy appends it to the network hash so each
  *   device behind a shared network gets its own bucket.
+ * - `req.tlsFingerprint` — set only when `tlsFingerprintHeader` is configured and
+ *   the request carries a well-shaped value in that header. It names the
+ *   client's TLS stack, so every client on the same build shares it.
  *
  * Install AFTER the body parsers (so the body fingerprint is readable) and
  * before `requestContext` (so the snapshot into ALS already carries them).
  * Never rejects — a missing/invalid device fingerprint is simply absent.
  */
-export const createFingerprintMiddleware = (): RequestHandler => {
+export const createFingerprintMiddleware = (
+    opts: { tlsFingerprintHeader?: string } = {},
+): RequestHandler => {
+    const tlsHeader = opts.tlsFingerprintHeader?.toLowerCase();
     return (req, _res, next) => {
         req.networkFingerprint = computeNetworkFingerprint(req);
         req.deviceFingerprint = readDeviceFingerprint(req);
+        req.tlsFingerprint = readTlsFingerprint(req, tlsHeader);
         next();
     };
 };

src/backend/core/http/middleware/largeJsonBody.ts

@@ -20,6 +20,7 @@
 import type { Request, RequestHandler, Response } from 'express';
 import { actorUid } from '../../actor';
 import { HttpError } from '../HttpError';
+import { onResponseClose } from '../onResponseClose';
 import { createJsonBodyParser } from './jsonBody';
 
 /** The global JSON parser's limit, which applies before auth. */
@@ -152,7 +153,7 @@ export const createLargeJsonBodyParser = ({
         }
         inFlightBytes += declared;
         inFlightByUser.set(user, userBytes + declared);
-        res.once('close', () => {
+        onResponseClose(res, () => {
             inFlightBytes -= declared;
             const left = (inFlightByUser.get(user) ?? 0) - declared;
             if (left > 0) inFlightByUser.set(user, left);

src/backend/core/http/middleware/originGate.ts

@@ -26,7 +26,7 @@ import type { IConfig } from '../../../types';
  * Browsers send none of those variations, but config values do — a hand-edited
  * `origin` with a trailing slash would otherwise reject every real request.
  */
-const normalizeOrigin = (raw: string | undefined): string =>
+export const normalizeOrigin = (raw: string | undefined): string =>
     (raw ?? '').trim().replace(/\/+$/, '').toLowerCase();
 
 /**

src/backend/core/http/middleware/rateLimit.js

@@ -20,9 +20,8 @@
 
 import crypto from 'node:crypto';
 import { withSpan } from '../../../util/span.js';
-import { Context } from '../../context.js';
-import { CallSlot } from '../callSlot.js';
 import { HttpError } from '../HttpError.js';
+import { onResponseClose } from '../onResponseClose.js';
 import { subscriptionOverride } from '../../../services/metering/consts.js';
 
 /**
@@ -414,8 +413,8 @@ async function peekKv(kv, key, limit) {
 //
 // Lifecycle is the wedge between rate and concurrent limits: rate just
 // records a tick, concurrent has to track "still in flight" → "done"
-// across a request boundary. The route middleware and the driver controller
-// both hold the slot in a `CallSlot`.
+// across a request boundary. The route middleware hooks `res.finish` /
+// `res.close`; the driver helper wraps the invocation in `try/finally`.
 
 // Orphan TTL safety nets — used when the process dies between acquire
 // and release. Memory backend doesn't need one (the process is also
@@ -1139,9 +1138,8 @@ async function resolveSubscriptionLimit(req, opts) {
  * bySubscription: { user_free: 2, unlimited: 50 } } } { concurrent: { limit:
  * 10, backend: 'redis', scope: 'expensive-op' } }
  *
- * On accept, holds the slot until the response is done and any work the handler
- * registered through `keepConcurrencySlot` has settled, so work that outlives a
- * caller who hung up still counts. On reject, 429 with the same
+ * On accept, schedules release on `res.finish` / `res.close` so even aborted
+ * requests give their slot back. On reject, 429 with the same
  * `too_many_requests` legacyCode as the rate gate (clients already handle that
  * branch). Fails open on backend error.
  */
@@ -1167,11 +1165,13 @@ export function concurrencyGate(opts) {
             return;
         }
 
-        const held = new CallSlot();
-        held.attach(guardSlot(slot, '[concurrent]'), res);
-        if (Context.current()) {
-            Context.set('keepConcurrencySlot', (work) => held.keep(work));
-        }
+        // `finish` (response sent) and `close` (connection closed, possibly
+        // aborted before finish) can both fire; the guard releases once.
+        const { release } = guardSlot(slot, '[concurrent]');
+        const onDone = () => void release();
+        res.once('finish', onDone);
+        onResponseClose(res, onDone);
+        if (res.destroyed || res.writableFinished) onDone();
         next();
     };
 }

src/backend/core/http/middleware/rateLimit.test.js

@@ -29,7 +29,6 @@ import {
     vi,
 } from 'vitest';
 import { EventEmitter } from 'node:events';
-import { Context, runWithContext } from '../../context.js';
 import { isHttpError } from '../HttpError.js';
 import { setupTestServer } from '../../../testUtil.ts';
 import {
@@ -978,27 +977,6 @@ describe('concurrencyGate — memory backend', () => {
         expect(isHttpError(third.err)).toBe(true);
     });
 
-    it('holds the slot past the response for work the handler keeps', async () => {
-        const opts = { limit: 1, key: 'ip', scope: 'cg-kept-work' };
-        let finishWork;
-        const work = new Promise((resolve) => {
-            finishWork = resolve;
-        });
-        const res = makeRes();
-        await runWithContext({}, async () => {
-            await concurrencyGate(opts)(baseReq(), res, () => {
-                Context.get('keepConcurrencySlot')(work);
-            });
-        });
-        res.emit('close');
-        await new Promise((r) => setImmediate(r));
-        expect((await runGate(opts, baseReq())).err?.statusCode).toBe(429);
-
-        finishWork();
-        await new Promise((r) => setImmediate(r));
-        expect((await runGate(opts, baseReq())).err).toBeUndefined();
-    });
-
     it('different keys do not share slots', async () => {
         const opts = { limit: 1, key: 'ip', scope: 'cg-keyed' };
         const a = await runGate(opts, baseReq({ ip: '1.1.1.1' }));

src/backend/core/http/onResponseClose.test.ts

@@ -0,0 +1,105 @@
+/*
+ * Copyright (C) 2024-present Puter Technologies Inc.
+ *
+ * This file is part of Puter.
+ *
+ * Puter is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published
+ * by the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <https://www.gnu.org/licenses/>.
+ */
+
+import { EventEmitter } from 'node:events';
+import { createServer, get, type Server } from 'node:http';
+import type { AddressInfo } from 'node:net';
+import { afterEach, describe, expect, it, vi } from 'vitest';
+import { onResponseClose } from './onResponseClose.js';
+
+let server: Server | undefined;
+afterEach(async () => {
+    server?.closeAllConnections();
+    await new Promise((resolve) => server?.close(resolve) ?? resolve(null));
+    server = undefined;
+});
+
+describe('onResponseClose', () => {
+    it('shares one close listener and runs every callback once, in order', () => {
+        const res = new EventEmitter();
+        const calls: string[] = [];
+        onResponseClose(res, () => calls.push('a'));
+        onResponseClose(res, () => calls.push('b'));
+        onResponseClose(res, () => calls.push('c'));
+        expect(res.listenerCount('close')).toBe(1);
+
+        res.emit('close');
+        res.emit('close');
+        expect(calls).toEqual(['a', 'b', 'c']);
+        expect(res.listenerCount('close')).toBe(0);
+    });
+
+    it('keeps responses apart', () => {
+        const first = new EventEmitter();
+        const second = new EventEmitter();
+        const onFirst = vi.fn();
+        const onSecond = vi.fn();
+        onResponseClose(first, onFirst);
+        onResponseClose(second, onSecond);
+
+        first.emit('close');
+        expect(onFirst).toHaveBeenCalledOnce();
+        expect(onSecond).not.toHaveBeenCalled();
+    });
+
+    it('runs the rest when one callback throws, and still surfaces the error', async () => {
+        const res = new EventEmitter();
+        const after = vi.fn();
+        const failure = new Error('callback broke');
+        const uncaught = vi.fn();
+        const listeners = process.listeners('uncaughtException');
+        process.removeAllListeners('uncaughtException');
+        process.on('uncaughtException', uncaught);
+        try {
+            onResponseClose(res, () => {
+                throw failure;
+            });
+            onResponseClose(res, after);
+            res.emit('close');
+            expect(after).toHaveBeenCalledOnce();
+            await vi.waitFor(() =>
+                expect(uncaught).toHaveBeenCalledWith(
+                    failure,
+                    expect.anything(),
+                ),
+            );
+        } finally {
+            process.off('uncaughtException', uncaught);
+            for (const listener of listeners)
+                process.on('uncaughtException', listener);
+        }
+    });
+
+    it('fires when a client hangs up on a real response', async () => {
+        const closed = new Promise<boolean>((resolve) => {
+            server = createServer((_req, res) => {
+                onResponseClose(res, () => resolve(res.writableFinished));
+                res.write('partial');
+            });
+        });
+        await new Promise<void>((resolve) => server!.listen(0, resolve));
+        const { port } = server!.address() as AddressInfo;
+        const req = get(`http://127.0.0.1:${port}/`, (res) => {
+            res.once('data', () => req.destroy());
+        });
+        req.on('error', () => {});
+
+        expect(await closed).toBe(false);
+    });
+});

src/backend/core/http/onResponseClose.ts

@@ -0,0 +1,57 @@
+/*
+ * Copyright (C) 2024-present Puter Technologies Inc.
+ *
+ * This file is part of Puter.
+ *
+ * Puter is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published
+ * by the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <https://www.gnu.org/licenses/>.
+ */
+
+/** Only what's used here, so test doubles and express's Response both fit. */
+interface ClosingResponse {
+    once(event: 'close', listener: () => void): unknown;
+}
+
+const callbacksByResponse = new WeakMap<ClosingResponse, Array<() => void>>();
+
+/**
+ * Runs `callback` once when the response closes. Every callback on a response
+ * shares one `close` listener, since Node warns past ten and request tracing
+ * adds its own. Callbacks run in the order added; one that throws doesn't stop
+ * the rest.
+ */
+export const onResponseClose = (
+    res: ClosingResponse,
+    callback: () => void,
+): void => {
+    const callbacks = callbacksByResponse.get(res);
+    if (callbacks) {
+        callbacks.push(callback);
+        return;
+    }
+    const pending = [callback];
+    callbacksByResponse.set(res, pending);
+    res.once('close', () => {
+        for (const run of pending.splice(0)) {
+            try {
+                run();
+            } catch (e) {
+                // Surfaces as an uncaught error, as it would from a listener
+                // of its own.
+                queueMicrotask(() => {
+                    throw e;
+                });
+            }
+        }
+    });
+};

src/backend/core/http/responseCloseListeners.http.test.ts

@@ -0,0 +1,147 @@
+/*
+ * Copyright (C) 2024-present Puter Technologies Inc.
+ *
+ * This file is part of Puter.
+ *
+ * Puter is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published
+ * by the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <https://www.gnu.org/licenses/>.
+ */
+
+import { afterAll, beforeAll, describe, expect, it } from 'vitest';
+import {
+    RESPONSE_CLOSE_LISTENER_BUDGET,
+    setupPuterTestEnv,
+    trackResponseCloseListeners,
+    type PuterTestEnv,
+} from '../../testUtil.js';
+
+/**
+ * The routes that hold a response open the longest, where middleware, gates and
+ * the handler each want to hear when it closes.
+ */
+describe('close listeners per response over HTTP', () => {
+    let env: PuterTestEnv;
+    let filePath = '';
+
+    beforeAll(async () => {
+        env = await setupPuterTestEnv({
+            meteringEnforcement: { subscriptions: false },
+        } as never);
+        const { username } = env.users.user;
+        const user = await env.server.stores.user.getByUsername(username);
+        filePath = `/${username}/Documents/close-listeners.txt`;
+        await env.server.services.fs.write(user!.id, {
+            fileMetadata: {
+                path: filePath,
+                size: 4096,
+                contentType: 'text/plain',
+            },
+            fileContent: Buffer.alloc(4096, 97),
+        });
+    }, 120_000);
+
+    afterAll(async () => {
+        await env?.shutdown();
+    });
+
+    const streamedChat = {
+        model: 'fake',
+        stream: true,
+        messages: [{ role: 'user', content: 'hi' }],
+        max_tokens: 16,
+    };
+    const post = (path: string, token: string, body: unknown) =>
+        fetch(new URL(path, env.apiOrigin), {
+            method: 'POST',
+            headers: {
+                Authorization: `Bearer ${token}`,
+                'Content-Type': 'application/json',
+                Origin: env.apiOrigin,
+            },
+            body: JSON.stringify(body),
+        });
+
+    const routes: Array<[string, string, () => Promise<Response>]> = [
+        [
+            'a streamed driver call',
+            '/drivers/call',
+            () =>
+                post('/drivers/call', env.users.user.token, {
+                    interface: 'puter-chat-completion',
+                    service: 'ai-chat',
+                    method: 'complete',
+                    args: streamedChat,
+                }),
+        ],
+        [
+            'an Anthropic-style streamed completion',
+            '/puterai/anthropic/v1/messages',
+            () =>
+                post(
+                    '/puterai/anthropic/v1/messages',
+                    env.users.user.apiToken,
+                    streamedChat,
+                ),
+        ],
+        [
+            'an OpenAI-style streamed completion',
+            '/puterai/openai/v1/chat/completions',
+            () =>
+                post(
+                    '/puterai/openai/v1/chat/completions',
+                    env.users.user.apiToken,
+                    streamedChat,
+                ),
+        ],
+        [
+            'a file read',
+            '/read',
+            () => {
+                const url = new URL('/read', env.apiOrigin);
+                url.searchParams.set('file', filePath);
+                return fetch(url, {
+                    headers: {
+                        Authorization: `Bearer ${env.users.user.token}`,
+                    },
+                });
+            },
+        ],
+        [
+            'a v2 file read',
+            '/fs/read',
+            () => {
+                const url = new URL('/fs/read', env.apiOrigin);
+                url.searchParams.set('path', filePath);
+                return fetch(url, {
+                    headers: {
+                        Authorization: `Bearer ${env.users.user.token}`,
+                    },
+                });
+            },
+        ],
+    ];
+
+    it.each(routes)('%s stays within the budget', async (_name, path, call) => {
+        const tracker = trackResponseCloseListeners();
+        try {
+            const response = await call();
+            expect(response.status).toBe(200);
+            await response.arrayBuffer();
+        } finally {
+            tracker.restore();
+        }
+        const peak = tracker.peak((req) => req.url?.split('?')[0] === path);
+        expect(peak).toBeGreaterThan(0);
+        expect(peak).toBeLessThanOrEqual(RESPONSE_CLOSE_LISTENER_BUDGET);
+    });
+});

src/backend/core/http/routeLifecycle.ts

@@ -21,6 +21,7 @@ import type { Request, RequestHandler } from 'express';
 import { actorUid } from '../actor';
 import type { EventClient } from '../../clients/event/EventClient';
 import type { LifecycleRejectStatus } from '../../clients/event/types';
+import { onResponseClose } from './onResponseClose';
 import type { RouteMethod } from './types';
 
 type RoutePathValue = string | RegExp | Array<string | RegExp> | undefined;
@@ -179,7 +180,7 @@ export const createRouteLifecycleMiddleware = (
             );
         };
         res.once('finish', () => settle(false));
-        res.once('close', () => settle(!res.writableFinished));
+        onResponseClose(res, () => settle(!res.writableFinished));
         next();
     };
 };

src/backend/core/http/types.ts

@@ -263,10 +263,10 @@ export interface RouteOptions {
     rateLimit?: RouteRateLimit | RouteRateLimit[];
 
     /**
-     * Cap simultaneous in-flight requests per key; a slot is released once the
-     * response is done and any work the handler kept it for has settled.
-     * `bySubscription` overrides `limit` per `SubscriptionPolicy.id` when
-     * metering is wired in. `key`, `scope`, `backend` as in `rateLimit`.
+     * Cap simultaneous in-flight requests per key; slots are released on
+     * response finish or close. `bySubscription` overrides `limit` per
+     * `SubscriptionPolicy.id` when metering is wired in. `key`, `scope`,
+     * `backend` as in `rateLimit`.
      */
     concurrent?: {
         limit: number;

src/backend/drivers/ai/interfaces/chat/ChatCompletionDriver.availability.http.test.ts

@@ -129,17 +129,6 @@ describe('model availability through real HTTP', () => {
             meteringEnforcement: { subscriptions: false },
         } as never);
     }, 120000);
-    it('serves qualified chat model names through HTTP', async () => {
-        const response = await fetch(
-            new URL('/puterai/chat/models', env.apiOrigin),
-        );
-        expect(response.status).toBe(200);
-        const { models } = (await response.json()) as { models: string[] };
-        expect(models).toContain('openrouter:vendor/availability-test');
-        expect(models.every((id) => /^[^:/]+:[^/]+\/.+/.test(id))).toBe(true);
-        expect(models).not.toContain('vendor/availability-test');
-    });
-
     afterAll(async () => {
         await env?.shutdown();
         upstream?.closeAllConnections();

src/backend/drivers/ai/interfaces/chat/ChatCompletionDriver.concurrency.http.test.ts

@@ -1,225 +0,0 @@
-/*
- * Copyright (C) 2024-present Puter Technologies Inc.
- *
- * This file is part of Puter.
- *
- * Puter is free software: you can redistribute it and/or modify
- * it under the terms of the GNU Affero General Public License as published
- * by the Free Software Foundation, either version 3 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU Affero General Public License for more details.
- *
- * You should have received a copy of the GNU Affero General Public License
- * along with this program.  If not, see <https://www.gnu.org/licenses/>.
- */
-
-import { createServer, type Server, type ServerResponse } from 'node:http';
-import { afterAll, beforeAll, describe, expect, it, vi } from 'vitest';
-import { setupPuterTestEnv, type PuterTestEnv } from '../../../../testUtil.js';
-
-/**
- * A completion runs on to its usage report after the caller hangs up, so it
- * keeps counting against the caller's concurrency cap until it settles.
- */
-describe('chat concurrency over HTTP after the caller hangs up', () => {
-    let upstream: Server;
-    let env: PuterTestEnv;
-    let onRequest: (res: ServerResponse, body: string) => void;
-
-    beforeAll(async () => {
-        upstream = createServer((req, res) => {
-            if (req.url === '/api/tags') {
-                res.setHeader('Content-Type', 'application/json');
-                res.end(JSON.stringify({ models: [{ name: 'slot-test' }] }));
-                return;
-            }
-            let body = '';
-            req.setEncoding('utf8');
-            req.on('data', (chunk: string) => {
-                body += chunk;
-            });
-            req.once('end', () => onRequest(res, body));
-        });
-        await new Promise<void>((resolve) =>
-            upstream.listen(0, '127.0.0.1', resolve),
-        );
-        const port = (upstream.address() as { port: number }).port;
-        env = await setupPuterTestEnv({
-            providers: {
-                ollama: { apiBaseUrl: `http://127.0.0.1:${port}` },
-            },
-            meteringEnforcement: { subscriptions: false },
-        } as never);
-        await vi.waitFor(async () => {
-            expect(
-                (await env.server.drivers.aiChat.models()).some(
-                    (model) => model.id === 'ollama:ollama/slot-test',
-                ),
-            ).toBe(true);
-        });
-    }, 120_000);
-
-    afterAll(async () => {
-        await env?.shutdown();
-        upstream?.closeAllConnections();
-        if (upstream)
-            await new Promise<void>((resolve) =>
-                upstream.close(() => resolve()),
-            );
-    });
-
-    const call = (route: string, stream: boolean, signal?: AbortSignal) => {
-        const args = {
-            model: 'ollama:ollama/slot-test',
-            stream,
-            messages: [{ role: 'user', content: 'hi' }],
-            max_tokens: 16,
-        };
-        return fetch(new URL(route, env.apiOrigin), {
-            method: 'POST',
-            headers: {
-                'Content-Type': 'application/json',
-                Authorization: `Bearer ${env.users.user.apiToken}`,
-                Origin: env.apiOrigin,
-            },
-            body: JSON.stringify(
-                route === '/drivers/call'
-                    ? {
-                          interface: 'puter-chat-completion',
-                          method: 'complete',
-                          args,
-                      }
-                    : args,
-            ),
-            signal,
-        });
-    };
-
-    const usage = { prompt_tokens: 3, completion_tokens: 5, total_tokens: 8 };
-    const chunk = (delta: object, finish: string | null, extra = {}) =>
-        `data: ${JSON.stringify({
-            id: 'chatcmpl_1',
-            object: 'chat.completion.chunk',
-            choices: [{ index: 0, delta, finish_reason: finish }],
-            ...extra,
-        })}\n\n`;
-
-    const finishAnswer = (res: ServerResponse, body: string) => {
-        if (!(JSON.parse(body) as { stream?: boolean }).stream) {
-            res.setHeader('Content-Type', 'application/json');
-            res.end(
-                JSON.stringify({
-                    id: 'chatcmpl_1',
-                    object: 'chat.completion',
-                    created: 0,
-                    model: 'slot-test',
-                    choices: [
-                        {
-                            index: 0,
-                            message: { role: 'assistant', content: 'done' },
-                            finish_reason: 'stop',
-                        },
-                    ],
-                    usage,
-                }),
-            );
-            return;
-        }
-        if (!res.headersSent)
-            res.setHeader('Content-Type', 'text/event-stream');
-        res.end(
-            `${chunk({ content: 'done' }, 'stop')}data: ${JSON.stringify({
-                id: 'chatcmpl_1',
-                object: 'chat.completion.chunk',
-                choices: [],
-                usage,
-            })}\n\ndata: [DONE]\n\n`,
-        );
-    };
-
-    /** Starts a completion, waits until it is generating, then hangs up. */
-    const abandon = async (route: string, stream: boolean) => {
-        let received!: () => void;
-        const reached = new Promise<void>((resolve) => {
-            received = resolve;
-        });
-        const pending: { res: ServerResponse; body: string }[] = [];
-        onRequest = (res, body) => {
-            pending.push({ res, body });
-            if (stream) {
-                res.setHeader('Content-Type', 'text/event-stream');
-                res.write(chunk({ role: 'assistant', content: 'partial' }, null));
-            }
-            received();
-        };
-        const abort = new AbortController();
-        const request = call(route, stream, abort.signal);
-        const outcome = await Promise.race([
-            reached.then(() => 'generating' as const),
-            request.then((response) => response),
-        ]);
-        if (outcome !== 'generating') {
-            return { refused: outcome, pending };
-        }
-        if (stream) {
-            const response = await request;
-            const reader = response.body!.getReader();
-            let text = '';
-            while (!text.includes('partial')) {
-                const { value, done } = await reader.read();
-                if (done) break;
-                text += new TextDecoder().decode(value);
-            }
-        }
-        abort.abort();
-        await request.catch(() => undefined);
-        return { refused: undefined, pending };
-    };
-
-    it.each([
-        { route: '/drivers/call', stream: true },
-        { route: '/drivers/call', stream: false },
-        { route: '/puterai/openai/v1/chat/completions', stream: true },
-        { route: '/puterai/openai/v1/chat/completions', stream: false },
-        { route: '/puterai/anthropic/v1/messages', stream: true },
-    ])(
-        'counts an abandoned completion until it settles: $route (stream=$stream)',
-        async ({ route, stream }) => {
-            const running: { res: ServerResponse; body: string }[] = [];
-            let refused: Response | undefined;
-            try {
-                for (let i = 0; i < 25 && !refused; i++) {
-                    const attempt = await abandon(route, stream);
-                    running.push(...attempt.pending);
-                    refused = attempt.refused;
-                    // Let the server see the hang-up before the next call.
-                    await new Promise((resolve) => setTimeout(resolve, 50));
-                }
-                expect(refused?.status).toBe(429);
-                expect(await refused!.text()).toContain(
-                    'Too many concurrent requests',
-                );
-                expect(running.length).toBeGreaterThan(0);
-
-                for (const { res, body } of running.splice(0))
-                    finishAnswer(res, body);
-                onRequest = finishAnswer;
-                await expect
-                    .poll(async () => (await call(route, false)).status, {
-                        timeout: 15_000,
-                    })
-                    .toBe(200);
-            } finally {
-                for (const { res, body } of running) finishAnswer(res, body);
-                upstream.closeAllConnections();
-                // The next case starts with a fresh rate-limit window.
-                await new Promise((resolve) => setTimeout(resolve, 10_100));
-            }
-        },
-        60_000,
-    );
-});

src/backend/drivers/ai/interfaces/chat/ChatCompletionDriver.edges.test.ts

@@ -1231,41 +1231,4 @@ describe('ChatCompletionDriver streaming failure handling', () => {
         expect(events).toContainEqual({ type: 'text', text: 'hello' });
         expect(cleanup).toHaveBeenCalledTimes(1);
     });
-
-    it('frees the concurrency slot before the provider cleanup hook finishes', async () => {
-        let finishCleanup!: () => void;
-        const cleanup = vi.fn(
-            () =>
-                new Promise<void>((resolve) => {
-                    finishCleanup = resolve;
-                }),
-        );
-        vi.spyOn(FakeChatProvider.prototype, 'complete').mockResolvedValue({
-            stream: true,
-            init_chat_stream: async ({
-                chatStream,
-            }: {
-                chatStream: { end: (usage?: Record<string, number>) => void };
-            }) => {
-                chatStream.end({ input_tokens: 1, output_tokens: 1 });
-            },
-            finally_fn: cleanup,
-        } as never);
-
-        const kept: Promise<unknown>[] = [];
-        const result = (await withTestActor(() => {
-            Context.set('keepConcurrencySlot', (work) => kept.push(work));
-            return fakeOnlyDriver.complete({
-                model: 'fake',
-                messages: [{ role: 'user', content: 'hi' }],
-                stream: true,
-            } as never);
-        })) as unknown as { stream: Readable };
-        await collect(result.stream);
-
-        expect(kept).toHaveLength(1);
-        await kept[0];
-        await vi.waitFor(() => expect(cleanup).toHaveBeenCalledTimes(1));
-        finishCleanup();
-    });
 });

src/backend/drivers/ai/interfaces/chat/ChatCompletionDriver.fuzz.test.ts

@@ -31,15 +31,14 @@ import { setupTestServer } from '../../../../testUtil.js';
 import { kv } from '../../../../util/kvSingleton.js';
 import { isDriverStreamResult } from '../../../meta.js';
 import {
-    caseRandom,
     corruptReply,
     Findings,
     fuzzCases,
     hostileValue,
     mutate,
     pick,
-    pickStable,
     type Random,
+    seededRandom,
     unclassifiedFailure,
     upstream,
     type UpstreamReply,
@@ -592,11 +591,7 @@ afterAll(async () => {
 });
 
 const pickRoute = (random: Random) =>
-    pickStable(
-        random,
-        routes.get(pickStable(random, [...routes.keys()], (name) => name))!,
-        (route) => route.id,
-    );
+    pick(random, routes.get(pick(random, [...routes.keys()]))!);
 
 /** Runs one completion to its end, stream included, and returns its failure. */
 const complete = async (args: Record<string, unknown>): Promise<unknown> => {
@@ -640,11 +635,10 @@ describe('ChatCompletionDriver seeded fuzzing', () => {
         upstream.respond = healthyUpstream;
         upstream.unanswered.clear();
         clearUnhealthyRoutes();
+        const random = seededRandom(0xc4a7);
         const findings = new Findings();
-        let answered = 0;
 
         for (let index = 0; index < fuzzCases(3000); index++) {
-            const random = caseRandom(0xc4a7, index);
             const route = pickRoute(random);
             const args: Record<string, unknown> = {
                 ...mutate(pick(random, CORPUS), random, OPTION_NAMES),
@@ -658,7 +652,6 @@ describe('ChatCompletionDriver seeded fuzzing', () => {
             }
             const label = `${route.provider}:${route.id}`;
             const error = await complete(args);
-            if (!error) answered++;
             const unclassified = error && unclassifiedFailure(error);
             if (unclassified) findings.add(unclassified, index, label, args);
             else if (isHttpError(error) && error.statusCode >= 500) {
@@ -684,16 +677,13 @@ describe('ChatCompletionDriver seeded fuzzing', () => {
             findings.add(`no upstream fixture for ${request}`, -1, '', {});
         }
         findings.assertNone();
-        // A run where nothing is answered has stopped reaching the providers.
-        expect(answered).toBeGreaterThan(0);
     }, 600_000);
 
     it('ends every request with an answer or a classified error when the upstream misbehaves', async () => {
+        const random = seededRandom(0x0b5e);
         const findings = new Findings();
-        const askedBefore = upstream.asked;
 
         for (let index = 0; index < fuzzCases(2000); index++) {
-            const random = caseRandom(0x0b5e, index);
             const route = pickRoute(random);
             const replies: unknown[] = [];
             upstream.respond = (request) => {
@@ -721,6 +711,5 @@ describe('ChatCompletionDriver seeded fuzzing', () => {
             clearUnhealthyRoutes();
         }
         findings.assertNone();
-        expect(upstream.asked).toBeGreaterThan(askedBefore);
     }, 600_000);
 });

src/backend/drivers/ai/interfaces/chat/ChatCompletionDriver.test.ts

@@ -207,50 +207,6 @@ describe('ChatCompletionDriver model map', () => {
         d.onServerShutdown();
     });
 
-    it('lists qualified model ids without publishing aliases as models', async () => {
-        const ids = [
-            'kimi/kimi-3',
-            'ollama:plain-model',
-            'ollama:vendor/qualified',
-            'vendor/tagged:free',
-        ];
-        vi.spyOn(OllamaChatProvider.prototype, 'models').mockResolvedValue(
-            ids.map((id) => ({
-                id,
-                aliases: ['legacy-' + id],
-                costs_currency: 'usd-cents',
-                costs: {},
-                max_tokens: 8192,
-            })),
-        );
-        const d = newDriver({ ollama: {} });
-        await d.onServerStart();
-        try {
-            expect(
-                (await d.list()).filter(
-                    (id) => !['abuse', 'costly', 'fake'].includes(id),
-                ),
-            ).toEqual([
-                'ollama:kimi/kimi-3',
-                'ollama:ollama/plain-model',
-                'ollama:vendor/qualified',
-                'ollama:vendor/tagged:free',
-            ]);
-            const models = (await d.models()).filter(
-                (model) => model.provider === 'ollama',
-            );
-            expect(models.map((model) => model.id).sort()).toEqual(
-                [...ids].sort(),
-            );
-            for (const model of models) {
-                expect(model.aliases).toContain(model.puterId);
-                expect(model.aliases).toContain('legacy-' + model.id);
-            }
-        } finally {
-            d.onServerShutdown();
-        }
-    });
-
     it('lists the catalog to a caller that asks while it is still loading', async () => {
         const d = newDriver({ ollama: { enabled: false } });
         d.onServerStart();

src/backend/drivers/ai/interfaces/chat/ChatCompletionDriver.ts

@@ -1308,17 +1308,8 @@ export class ChatCompletionDriver extends PuterDriver {
             }, HOLD_RENEW_INTERVAL_MS);
             renewHold.unref?.();
 
-            // The upstream runs on after a caller hangs up, so the request
-            // keeps its concurrency slot until the upstream is done, not until
-            // the cleanup after it.
-            let upstreamDone!: () => void;
-            const upstreamRunning = new Promise<void>((resolve) => {
-                upstreamDone = resolve;
-            });
-            Context.get('keepConcurrencySlot')?.(upstreamRunning);
-
-            // The stream writes happen async while the response is piped to
-            // the client.
+            // Fire-and-forget — the stream writes happen async while the
+            // response is being piped to the client.
             (async () => {
                 try {
                     await init({ chatStream });
@@ -1343,7 +1334,6 @@ export class ChatCompletionDriver extends PuterDriver {
                     }
                     passthrough.end();
                 } finally {
-                    upstreamDone();
                     clearInterval(renewHold);
                     clearTimeout(detachedLimit);
                     abortSignal?.removeEventListener('abort', detachStream);
@@ -2086,39 +2076,14 @@ export class ChatCompletionDriver extends PuterDriver {
                 // instance, so they are read and never written: normalizing
                 // the id or appending puterId in place would accumulate across
                 // instantiations. The bucket gets its own copy instead.
-                // Keep upstream ids for provider calls and publish a qualified
-                // routing name even when a catalog omits its vendor namespace.
-                const publicId = entry.puterId || entry.id;
-                // A colon after the first slash belongs to the model name, as
-                // in `vendor/model:free`, and names no service.
-                const serviceSeparator = publicId
-                    .split('/', 1)[0]!
-                    .indexOf(':');
-                const service =
-                    serviceSeparator < 0
-                        ? providerName
-                        : publicId.slice(0, serviceSeparator);
-                const modelPath =
-                    serviceSeparator < 0
-                        ? publicId
-                        : publicId.slice(serviceSeparator + 1);
-                const puterId =
-                    providerName === 'fake-chat'
-                        ? entry.puterId
-                        : `${service}:${modelPath.includes('/') ? modelPath : `${service}/${modelPath}`}`;
-                const aliases = puterId
-                    ? [
-                          ...new Set([
-                              ...(entry.aliases ?? []),
-                              ...(entry.puterId ? [entry.puterId] : []),
-                              puterId,
-                          ]),
-                      ]
-                    : entry.aliases;
+                const aliases =
+                    entry.puterId &&
+                    !(entry.aliases ?? []).includes(entry.puterId)
+                        ? [...(entry.aliases ?? []), entry.puterId]
+                        : entry.aliases;
                 const model = {
                     ...entry,
                     id: normalizeModelKey(entry.id),
-                    ...(puterId ? { puterId } : {}),
                 };
                 // Assigned only when the entry has names to carry: models()
                 // is serialized to the API, and an entry that declared no

src/backend/drivers/ai/interfaces/image/ImageGenerationDriver.test.ts

@@ -496,19 +496,6 @@ describe('ImageGenerationDriver.generate provider routing', () => {
         });
     });
 
-    it('ignores an input image type that is not a string when there is no input image', async () => {
-        openaiImagesGenerateMock.mockResolvedValueOnce({
-            data: [{ url: 'https://oai/image.png' }],
-        });
-        await expect(withDriverName('ai-image', () =>
-            driver.generate({
-                model: 'gpt-image-2',
-                prompt: 'a landscape',
-                input_image_mime_type: 5,
-            } as never),
-        )).resolves.toBe('https://oai/image.png');
-    });
-
     it('rejects an excluded model regardless of caller casing', async () => {
         await expect(withActor(() =>
             driver.generate({

src/backend/drivers/ai/interfaces/image/ImageGenerationDriver.ts

@@ -177,14 +177,10 @@ export class ImageGenerationDriver extends PuterDriver {
         // settled here — once — before any provider runs.
         assertImagePrompt(args.prompt);
         assertInputImagesShape(args, 'image generation');
-        // The type hint describes an input image, so it is only read, and
-        // only has to be a string, when the request carries one.
-        const hasInputImage =
-            args.input_image != null || (args.input_images?.length ?? 0) > 0;
         for (const option of [
             'quality',
             'resolution',
-            ...(hasInputImage ? (['input_image_mime_type'] as const) : []),
+            'input_image_mime_type',
         ] as const) {
             if (args[option] != null && typeof args[option] !== 'string') {
                 throw new HttpError(400, `${option} must be a string`, {
@@ -197,7 +193,6 @@ export class ImageGenerationDriver extends PuterDriver {
         // caller's `args` is re-emitted by reference in `.after`/`.error`
         // payloads, so it is never mutated here.
         const { puter_output_path: puterOutputPath, ...request } = args;
-        if (!hasInputImage) delete request.input_image_mime_type;
 
         // Validate the output path early — before spending credits.
         const resolvedOutputPath = puterOutputPath

src/backend/drivers/ai/interfaces/mediaDrivers.fuzz.test.ts

@@ -37,7 +37,6 @@ import {
     fuzzCases,
     hostileValue,
     pick,
-    pickStable,
     type Random,
     upstream,
 } from '../shared/fuzzTestUtil.js';
@@ -363,12 +362,9 @@ describe.each(TARGETS)('$name driver seeded fuzzing', (target) => {
 
         const pinned = (random: Random) => {
             const route: Record<string, unknown> = {
-                ...pickStable(
+                ...pick(
                     random,
-                    byProvider.get(
-                        pickStable(random, target.providers, (name) => name),
-                    )!,
-                    (candidate) => String(candidate.model),
+                    byProvider.get(pick(random, target.providers))!,
                 ),
             };
             // Now and then the route itself is the malformed part.
@@ -378,7 +374,6 @@ describe.each(TARGETS)('$name driver seeded fuzzing', (target) => {
             }
             return route;
         };
-        const askedBefore = upstream.asked;
         const findings = await fuzzCalls({
             seed: 0x3ed1a,
             cases: fuzzCases(1000),
@@ -397,7 +392,5 @@ describe.each(TARGETS)('$name driver seeded fuzzing', (target) => {
             },
         });
         findings.assertNone();
-        // Some of the requests have to get as far as a vendor.
-        expect(upstream.asked).toBeGreaterThan(askedBefore);
     }, 600_000);
 });

src/backend/drivers/ai/interfaces/ocr/OCRDriver.ts

@@ -135,11 +135,9 @@ export class OCRDriver extends PuterDriver {
         const metering = this.#aiMetering;
 
         const textract = providers['aws-textract'] as
-            | Record<string, unknown>
-            | undefined;
+            Record<string, unknown> | undefined;
         const textractAws = (textract?.aws ?? textract) as
-            | Record<string, unknown>
-            | undefined;
+            Record<string, unknown> | undefined;
         const textractCredentials = awsCredentialsFrom(textractAws);
         if (textractCredentials) {
             this.#textract = new TextractOcrProvider(metering, {

src/backend/drivers/ai/interfaces/speechToSpeech/VoiceChangerDriver.ts

@@ -115,11 +115,9 @@ export class VoiceChangerDriver extends PuterDriver {
                     apiKey,
                     apiBaseUrl: elevenlabs?.apiBaseUrl as string | undefined,
                     defaultVoiceId: elevenlabs?.defaultVoiceId as
-                        | string
-                        | undefined,
+                        string | undefined,
                     defaultModelId: elevenlabs?.speechToSpeechModelId as
-                        | string
-                        | undefined,
+                        string | undefined,
                 },
             ),
         );

src/backend/drivers/ai/interfaces/speechToText/SpeechToTextDriver.test.ts

@@ -340,22 +340,6 @@ describe('SpeechToTextDriver argument validation', () => {
         ).rejects.toMatchObject({ statusCode: 400 });
     });
 
-    it('throws 400 when prompt is a list instead of text', async () => {
-        const { actor } = await makeUser();
-        await expect(
-            withActor(actor, () =>
-                driver.transcribe({
-                    file: dataUrl(Buffer.from('a'), 'audio/mp3'),
-                    prompt: [null, 'names'],
-                } as never),
-            ),
-        ).rejects.toMatchObject({
-            statusCode: 400,
-            message: 'prompt must be a string',
-        });
-        expect(transcriptionsCreateMock).not.toHaveBeenCalled();
-    });
-
     it('rejects diarized_json, which gpt-transcribe does not produce', async () => {
         const { actor } = await makeUser();
         await expect(

src/backend/drivers/ai/interfaces/video/VideoGenerationDriver.test.ts

@@ -552,8 +552,6 @@ describe('VideoGenerationDriver.generate provider routing', () => {
             'kling-3.0',
             'kling-3.0-omni',
             'kling-3.0-turbo',
-            'kling-4.0',
-            'kling-4.0-flash',
             'kling-o1',
         ]);
     });

src/backend/drivers/ai/protocols/openaiCompatible/OpenAIUtil.js

@@ -558,8 +558,7 @@ export const splitCachedPrompt = (usage) => {
  * @param {{ writeTtl?: '5m' | '30m' }} [opts] The lifetime `cache_write_tokens`
  *   were written with: 5m unless the provider says otherwise (OpenAI's default
  *   entries live 30m).
- * @returns {import('../../interfaces/chat/types.js').UsageDetails
- *     | undefined}
+ * @returns {import('../../interfaces/chat/types.js').UsageDetails | undefined}
  */
 export const usageDetailsFromTrackedUsage = (
     trackedUsage,

src/backend/drivers/ai/providers/gemini/chat/GeminiChatProvider.integration.test.ts

@@ -113,6 +113,7 @@ describe.skipIf(skipUnlessEnv(ENV_VAR))(
                 expect(typeof text === 'string' && text.length > 0).toBe(true);
             },
         );
+
     },
 );
 
@@ -376,36 +377,3 @@ describe.skipIf(skipUnlessEnv(ENV_VAR))(
         );
     },
 );
-
-describe.skipIf(skipUnlessEnv(ENV_VAR))(
-    'Gemini 4 Carbon and Hydrogen (integration)',
-    () => {
-        it.each([
-            ['gemini-4-carbon', 'low'],
-            ['gemini-4-hydrogen', 'xhigh'],
-        ])(
-            'generates a native text completion from %s at %s effort',
-            { timeout: INTEGRATION_TEST_TIMEOUT_MS },
-            async (model, effort) => {
-                const provider = new GeminiChatProvider(
-                    { apiKey: optionalEnv(ENV_VAR)! },
-                    makeMeteringStub(),
-                );
-                const result = await withTestActor(() =>
-                    provider.complete({
-                        model,
-                        messages: [{ role: 'user', content: 'Reply with OK.' }],
-                        reasoning_effort: effort,
-                        max_tokens: 32768,
-                    } as ICompleteArguments),
-                );
-                const message = result as IChatMessageResult;
-                expect(typeof message.message.content).toBe('string');
-                expect(
-                    (message.message.content as string).trim().length,
-                ).toBeGreaterThan(0);
-                expectBilledOutput(message.usage, message.usageDetails);
-            },
-        );
-    },
-);

src/backend/drivers/ai/providers/gemini/chat/GeminiChatProvider.native.test.ts

@@ -114,8 +114,6 @@ const makeProvider = (
 const ARGON = GEMINI_MODELS.find((m) => m.id === 'gemini-4-argon')!;
 const LAUNCH = new Date('2026-10-14T12:00:00Z');
 const NATIVE_IDS = [
-    'gemini-4-hydrogen',
-    'gemini-4-carbon',
     'gemini-4-argon',
     'gemini-3.1-flash-lite',
     'gemini-3-flash-preview',
@@ -1618,119 +1616,3 @@ describe('geminiNativeClient deadline', () => {
         expect(isUpstreamTimeoutError(err)).toBe(false);
     });
 });
-
-describe('Gemini 4 Carbon', () => {
-    it.each([
-        ['2026-11-30T23:59:59.999Z', 200, 1000, 10],
-        ['2026-12-01T00:00:00.000Z', 400, 2000, 20],
-    ])(
-        'routes natively and meters thinking at %s',
-        async (date, input, output, cached) => {
-            generateContent.mockResolvedValueOnce(
-                textResponse('OK', { ...LIVE_USAGE, cachedContentTokenCount: 6 }),
-            );
-            await complete(
-                {
-                    model: 'google/gemini-4-carbon',
-                    temperature: 0.5,
-                    reasoning_effort: 'medium',
-                },
-                new Date(date),
-            );
-            expect(requestOf().model).toBe('gemini-4-carbon');
-            expect(requestOf().config).not.toHaveProperty('temperature');
-            expect(openAICreate).not.toHaveBeenCalled();
-            expect(recordSpy.mock.calls[0]![3]).toMatchObject({
-                prompt_tokens: 30 * input,
-                completion_tokens: 5 * output,
-                thinking_tokens: 69 * output,
-                cached_tokens: 6 * cached,
-            });
-        },
-    );
-
-    it.each([
-        ['no max_tokens', undefined],
-        ['a larger max_tokens', 65_536],
-    ])(
-        'caps a non-streamed request with %s at 32,768 output tokens',
-        async (_label, maxTokens) => {
-            generateContent.mockResolvedValueOnce(textResponse('OK'));
-            await complete({ model: 'gemini-4-carbon', max_tokens: maxTokens });
-            expect(requestOf().config.maxOutputTokens).toBe(32_768);
-        },
-    );
-
-    it('sends a stream the max_tokens it asked for', async () => {
-        generateContentStream.mockResolvedValueOnce(
-            asAsyncIterable([textResponse('OK')]),
-        );
-        await runStream({ model: 'gemini-4-carbon', max_tokens: 65_536 });
-        expect(requestOf(generateContentStream).config.maxOutputTokens).toBe(
-            65_536,
-        );
-    });
-});
-
-describe('Gemini 4 Hydrogen', () => {
-    const hydrogen = (args: Partial<ICompleteArguments>) =>
-        complete(
-            { model: 'gemini-4-hydrogen', ...args },
-            new Date('2026-12-01T00:00:00.000Z'),
-        );
-
-    it('routes natively and meters thinking at its own rates', async () => {
-        generateContent.mockResolvedValueOnce(
-            textResponse('OK', { ...LIVE_USAGE, cachedContentTokenCount: 6 }),
-        );
-        await hydrogen({ model: 'google/gemini-4-hydrogen', temperature: 0.5 });
-        expect(requestOf().model).toBe('gemini-4-hydrogen');
-        expect(requestOf().config).not.toHaveProperty('temperature');
-        expect(requestOf().config).not.toHaveProperty('thinkingConfig');
-        expect(openAICreate).not.toHaveBeenCalled();
-        expect(recordSpy.mock.calls[0]![2]).toBe('gemini:gemini-4-hydrogen');
-        expect(recordSpy.mock.calls[0]![3]).toMatchObject({
-            prompt_tokens: 30 * 150,
-            completion_tokens: 5 * 750,
-            thinking_tokens: 69 * 750,
-            cached_tokens: 6 * 7.5,
-        });
-    });
-
-    it.each([
-        ['low', 'LOW'],
-        ['medium', 'MEDIUM'],
-        ['high', 'HIGH'],
-        ['xhigh', 'XHIGH'],
-    ])('sends reasoning_effort %s as thinkingLevel %s', async (effort, level) => {
-        generateContent.mockResolvedValueOnce(textResponse('OK'));
-        await hydrogen({
-            reasoning_effort: effort as ICompleteArguments['reasoning_effort'],
-        });
-        expect(requestOf().config.thinkingConfig).toEqual({
-            thinkingLevel: level,
-        });
-    });
-
-    it.each(['none', 'minimal', 'max'])(
-        'rejects reasoning_effort %s before calling Google',
-        async (effort) => {
-            await expect(
-                hydrogen({
-                    reasoning_effort:
-                        effort as ICompleteArguments['reasoning_effort'],
-                }),
-            ).rejects.toMatchObject({
-                statusCode: 400,
-                legacyCode: 'bad_request',
-            });
-            expect(generateContent).not.toHaveBeenCalled();
-        },
-    );
-
-    it('sends a non-streamed request the max_tokens it asked for', async () => {
-        generateContent.mockResolvedValueOnce(textResponse('OK'));
-        await hydrogen({ max_tokens: 65_536 });
-        expect(requestOf().config.maxOutputTokens).toBe(65_536);
-    });
-});

src/backend/drivers/ai/providers/gemini/chat/GeminiChatProvider.ts

@@ -42,18 +42,22 @@ import {
 } from '../../../protocols/openaiCompatible/OpenAICompatProvider.js';
 import { usesGoogleGeminiEndpoint } from '../geminiRest.js';
 import { geminiNativeClient } from './geminiNative.js';
-import { GEMINI_MODELS, GEMINI_NATIVE_MODELS } from './models.js';
+import {
+    GEMINI_MODELS,
+    GEMINI_NATIVE_MODELS,
+    GEMINI_ARGON_NON_STREAM_MAX_TOKENS,
+} from './models.js';
 
 type GroundingContent = { grounding_metadata?: unknown };
 
-/** Some native models have a lower output cap when the response isn't streamed. */
+/** Unary Argon uses a lower output cap than streaming. */
 const nativeOutputCap = (
     args: ICompleteArguments,
     model: IChatModel,
 ): number | undefined =>
-    args.stream
-        ? undefined
-        : GEMINI_NATIVE_MODELS.get(model.id)?.unaryMaxTokens;
+    model.id === 'gemini-4-argon' && !args.stream
+        ? GEMINI_ARGON_NON_STREAM_MAX_TOKENS
+        : undefined;
 
 const COMPAT_BASE_URL =
     'https://generativelanguage.googleapis.com/v1beta/openai/';

src/backend/drivers/ai/providers/gemini/chat/geminiNative.ts

@@ -78,7 +78,7 @@ interface ChatParams {
     tool_choice?: ChatToolChoice;
     max_completion_tokens?: number;
     temperature?: number;
-    reasoning_effort?: 'minimal' | 'low' | 'medium' | 'high' | 'xhigh';
+    reasoning_effort?: 'minimal' | 'low' | 'medium' | 'high';
     stream?: boolean;
     service_tier?: string;
 }

src/backend/drivers/ai/providers/gemini/chat/models.ts

@@ -29,8 +29,6 @@ export interface GeminiNativeOptions {
     reasoningEfforts: ReadonlySet<ReasoningEffort>;
     /** False where Google ignores temperature, so it isn't sent. */
     temperature: boolean;
-    /** Output cap of a response that isn't streamed, where Google's is lower. */
-    unaryMaxTokens?: number;
 }
 
 const FLASH_NATIVE_OPTIONS: GeminiNativeOptions = {
@@ -38,25 +36,16 @@ const FLASH_NATIVE_OPTIONS: GeminiNativeOptions = {
     temperature: true,
 };
 
-// Google's model metadata reports the unary cap; streaming keeps the catalog cap.
-const ARGON_CARBON_NATIVE_OPTIONS: GeminiNativeOptions = {
-    reasoningEfforts: new Set(['low', 'medium', 'high']),
-    temperature: false,
-    unaryMaxTokens: 32_768,
-};
-
 /**
  * Served natively: the OpenAI-compatible endpoint omits these models' thinking
  * count from usage.
  */
 export const GEMINI_NATIVE_MODELS: ReadonlyMap<string, GeminiNativeOptions> =
     new Map([
-        ['gemini-4-argon', ARGON_CARBON_NATIVE_OPTIONS],
-        ['gemini-4-carbon', ARGON_CARBON_NATIVE_OPTIONS],
         [
-            'gemini-4-hydrogen',
+            'gemini-4-argon',
             {
-                reasoningEfforts: new Set(['low', 'medium', 'high', 'xhigh']),
+                reasoningEfforts: new Set(['low', 'medium', 'high']),
                 temperature: false,
             },
         ],
@@ -99,7 +88,7 @@ const FLASH_LIST_PRICING = listPricingFrom(
     FLASH_LAUNCH_COSTS,
 );
 
-const GEMINI_4_LAUNCH_COSTS: ModelCost = {
+const ARGON_LAUNCH_COSTS: ModelCost = {
     tokens: 1_000_000,
     prompt_tokens: 200,
     completion_tokens: 1000,
@@ -130,6 +119,9 @@ const PRO_LONG_CONTEXT_PRICING: ModelPricingOverride[] = [
     },
 ];
 
+// Google's unary model metadata reports this cap; streaming keeps the catalog cap.
+export const GEMINI_ARGON_NON_STREAM_MAX_TOKENS = 32_768;
+
 export const GEMINI_MODELS: IChatModel[] = [
     {
         puterId: 'google:google/gemini-4-argon',
@@ -151,66 +143,12 @@ export const GEMINI_MODELS: IChatModel[] = [
         input_cost_key: 'prompt_tokens',
         output_cost_key: 'completion_tokens',
         serviceSpeedRates: { fast: 1.8 },
-        costs: GEMINI_4_LAUNCH_COSTS,
+        costs: ARGON_LAUNCH_COSTS,
         pricingOverrides: listPricingFrom(
             '2026-12-01T00:00:00Z',
-            GEMINI_4_LAUNCH_COSTS,
+            ARGON_LAUNCH_COSTS,
         ),
     },
-    {
-        puterId: 'google:google/gemini-4-carbon',
-        id: 'gemini-4-carbon',
-        // The native adapter forwards only text and image parts.
-        modalities: {
-            input: ['text', 'image'],
-            output: ['text'],
-        },
-        open_weights: false,
-        tool_call: true,
-        release_date: '2026-11-10',
-        name: 'Gemini 4 Carbon',
-        aliases: ['google/gemini-4-carbon'],
-        // One window shared by input, thinking and output.
-        context: 1_048_576,
-        max_tokens: 65_536,
-        costs_currency: 'usd-cents',
-        input_cost_key: 'prompt_tokens',
-        output_cost_key: 'completion_tokens',
-        serviceSpeedRates: { fast: 1.8 },
-        costs: GEMINI_4_LAUNCH_COSTS,
-        pricingOverrides: listPricingFrom(
-            '2026-12-01T00:00:00Z',
-            GEMINI_4_LAUNCH_COSTS,
-        ),
-    },
-    {
-        puterId: 'google:google/gemini-4-hydrogen',
-        id: 'gemini-4-hydrogen',
-        // The native adapter forwards only text and image parts.
-        modalities: {
-            input: ['text', 'image'],
-            output: ['text'],
-        },
-        open_weights: false,
-        tool_call: true,
-        release_date: '2026-11-12',
-        name: 'Gemini 4 Hydrogen',
-        aliases: ['google/gemini-4-hydrogen'],
-        // One window shared by input, thinking and output.
-        context: 1_048_576,
-        max_tokens: 65_536,
-        costs_currency: 'usd-cents',
-        input_cost_key: 'prompt_tokens',
-        output_cost_key: 'completion_tokens',
-        serviceSpeedRates: { fast: 1.8 },
-        costs: {
-            tokens: 1_000_000,
-            prompt_tokens: 150,
-            completion_tokens: 750,
-            thinking_tokens: 750,
-            cached_tokens: 7.5,
-        },
-    },
     {
         puterId: 'google:google/gemini-3.5-flash',
         id: 'gemini-3.5-flash',

src/backend/drivers/ai/providers/gemini/video/GeminiVideoProvider.ts

@@ -193,8 +193,7 @@ interface Interaction {
 }
 
 type CountTokensPart =
-    | { text: string }
-    | { inlineData: { mimeType: string; data: string } };
+    { text: string } | { inlineData: { mimeType: string; data: string } };
 
 interface CountTokensRequest {
     contents: { parts: CountTokensPart[] }[];

src/backend/drivers/ai/providers/kling/video/KlingVideoProvider.integration.test.ts

@@ -39,10 +39,10 @@ const VIDEO_TIMEOUT_MS = 5 * 60 * 1000;
 describe.skipIf(skipUnlessEnv(ENV_VAR))(
     'KlingVideoProvider (integration)',
     () => {
-        it.each(['kling-2.5-turbo', 'kling-4.0', 'kling-4.0-flash'])(
-            'generates a video URL from %s',
+        it(
+            'generates a video URL from kling-2.5-turbo',
             { timeout: VIDEO_TIMEOUT_MS },
-            async (model) => {
+            async () => {
                 const provider = new KlingVideoProvider(
                     { apiKey: optionalEnv(ENV_VAR)! },
                     makeMeteringStub(),
@@ -51,7 +51,7 @@ describe.skipIf(skipUnlessEnv(ENV_VAR))(
 
                 const url = await withTestActor(() =>
                     provider.generate({
-                        model,
+                        model: 'kling-2.5-turbo',
                         prompt: 'a red ball rolls to the right',
                         resolution: '720p',
                         seconds: 5,

src/backend/drivers/ai/providers/kling/video/KlingVideoProvider.test.ts

@@ -37,7 +37,6 @@ import {
     type MockInstance,
 } from 'vitest';
 
-import { translateProviderError } from '../../../../../controllers/drivers/DriverController.js';
 import { Context } from '../../../../../core/context.js';
 import type { MeteringService } from '../../../../../services/metering/MeteringService.js';
 import { PuterServer } from '../../../../../server.js';
@@ -159,12 +158,10 @@ describe('KlingVideoProvider construction and catalog', () => {
         ).toThrow(/API key/i);
     });
 
-    it('defaults to Kling 3.0 Turbo and lists the supported models', () => {
+    it('defaults to Kling 3.0 Turbo and lists the six models', () => {
         const provider = makeProvider();
         expect(provider.getDefaultModel()).toBe('kling-3.0-turbo');
         expect(provider.models().map((m) => m.id)).toEqual([
-            'kling-4.0',
-            'kling-4.0-flash',
             'kling-3.0-turbo',
             'kling-3.0',
             'kling-3.0-omni',
@@ -747,112 +744,3 @@ describe('KlingVideoProvider after the caller leaves', () => {
         expect(incrementUsageSpy).not.toHaveBeenCalled();
     });
 });
-
-describe('Kling 4 generation', () => {
-    it.each([
-        ['kling-4.0', 30, '1080p', 23.55],
-        ['kling-4.0-flash', 20, '720p', 9.42],
-    ])(
-        'generates and bills %s at flagship rates',
-        async (model, seconds, size, cents) => {
-            mockTaskFlow(succeededTask());
-            await generate({
-                model,
-                seconds,
-                size,
-                prompt: 'a'.repeat(26_000),
-                generate_audio: true,
-            });
-            expect(sentUrl()).toBe(
-                `https://api-singapore.klingai.com/text-to-video/${model}`,
-            );
-            expect(sentBody().settings).toMatchObject({
-                duration: seconds,
-                resolution: size,
-            });
-            expect(sentBody().settings).not.toHaveProperty('audio');
-            expect(incrementUsageSpy.mock.calls[0]![3]).toBe(
-                Math.round(seconds * cents * 1_000_000),
-            );
-        },
-    );
-    it('routes Flash first frames to image-to-video', async () => {
-        mockTaskFlow(succeededTask());
-        await generate({
-            model: 'kling-4.0-flash',
-            input_reference: 'https://example.com/frame.png',
-        });
-        expect(sentUrl()).toBe(
-            'https://api-singapore.klingai.com/image-to-video/kling-4.0-flash',
-        );
-    });
-    it('rejects Flash last frames before submitting a task', async () => {
-        await expect(
-            generate({
-                model: 'kling-4.0-flash',
-                input_reference: 'https://example.com/frame.png',
-                last_frame: 'https://example.com/end.png',
-            }),
-        ).rejects.toMatchObject({ statusCode: 400 });
-        expect(fetchSpy).not.toHaveBeenCalled();
-    });
-    it('routes Kling 4 first and last frames to image-to-video', async () => {
-        mockTaskFlow(succeededTask());
-        await generate({
-            model: 'kling/kling-4.0',
-            input_reference: 'https://example.com/frame.png',
-            last_frame: 'https://example.com/end.png',
-            seconds: 3,
-        });
-        expect(sentUrl()).toBe(
-            'https://api-singapore.klingai.com/image-to-video/kling-4.0',
-        );
-        expect(sentBody().contents).toContainEqual({
-            type: 'last_frame',
-            url: 'https://example.com/end.png',
-        });
-        expect(sentBody().settings.duration).toBe(3);
-    });
-
-    it.each(['kling-4.0', 'kling-4.0-flash'])(
-        'rejects an oversized prompt on %s before submitting a task',
-        async (model) => {
-            await expect(
-                generate({ model, prompt: 'a'.repeat(26_001) }),
-            ).rejects.toMatchObject({ statusCode: 400 });
-            expect(fetchSpy).not.toHaveBeenCalled();
-        },
-    );
-
-    it("leaves Kling 4's lower non-English prompt limit to Kling", async () => {
-        fetchSpy.mockResolvedValueOnce(
-            jsonResponse(
-                {
-                    code: 1201,
-                    message:
-                        'prompt is too long: at most 26000 English characters or 11000 non-English characters',
-                },
-                400,
-            ),
-        );
-        const failure = await generate({
-            model: 'kling-4.0',
-            prompt: '海'.repeat(12_000),
-        }).catch((e) => e);
-        expect(sentBody().prompt).toHaveLength(12_000);
-        expect(translateProviderError(failure)).toMatchObject({
-            statusCode: 400,
-            legacyCode: 'upstream_bad_request',
-        });
-        expect(incrementUsageSpy).not.toHaveBeenCalled();
-    });
-
-    it.each(['kling-4.0', 'kling-4.0-flash'])(
-        'omits the unsupported audio switch on %s when audio is disabled',
-        async (model) => {
-            mockTaskFlow(succeededTask());
-            await generate({ model, generate_audio: false });
-            expect(sentBody().settings).not.toHaveProperty('audio');
-        },
-    );
-});

src/backend/drivers/ai/providers/kling/video/models.ts

@@ -62,44 +62,12 @@ const seconds = (def: number, min: number, max: number): number[] => [
 ];
 
 // Prices are per second of output from
-// https://kling.ai/document-api/pricing/base/video (1 unit = $0.14) and, for
-// Kling 4, https://kling.ai/dev/pricing (1 unit = $0.157). Cost keys are
-// `per-second:<resolution>`, with `:audio` for optional native audio and
+// https://kling.ai/document-api/pricing/base/video (1 unit = $0.14). Cost keys
+// are `per-second:<resolution>`, with `:audio` for native audio and
 // `:video-input` when a source video is edited.
 // `default-duration-per-video` is a 5-second clip at the default resolution,
 // for cross-provider sorting and display only.
 export const KLING_VIDEO_GENERATION_MODELS: IVideoModel[] = [
-    {
-        id: 'kling-4.0',
-        puterId: 'kling:kling/kling-4.0',
-        aliases: ['kling/kling-4.0'],
-        name: 'Kling 4.0',
-        costs_currency: 'usd-cents',
-        output_cost_key: 'default-duration-per-video',
-        costs: {
-            'per-second:720p': 15.7,
-            'per-second:1080p': 23.55,
-            'default-duration-per-video': 78.5,
-        },
-        durationSeconds: seconds(5, 3, 30),
-        dimensions: ['720p', '1080p'],
-        fps: [24],
-    },
-    {
-        id: 'kling-4.0-flash',
-        puterId: 'kling:kling/kling-4.0-flash',
-        aliases: ['kling/kling-4.0-flash'],
-        name: 'Kling 4.0 Flash',
-        costs_currency: 'usd-cents',
-        output_cost_key: 'default-duration-per-video',
-        costs: {
-            'per-second:720p': 9.42,
-            'default-duration-per-video': 47.1,
-        },
-        durationSeconds: seconds(5, 3, 20),
-        dimensions: ['720p'],
-        fps: [24],
-    },
     {
         id: 'kling-3.0-turbo',
         puterId: 'kling:kling/kling-3.0-turbo',
@@ -218,26 +186,6 @@ export const KLING_VIDEO_GENERATION_MODELS: IVideoModel[] = [
 // https://kling.ai/document-api/api/video and the capability map
 // https://kling.ai/document-api/guides/capability-map/video.
 export const KLING_VIDEO_SPECS: Record<string, KlingVideoSpec> = {
-    // Kling 4 takes 26,000 English characters but 11,000 non-English ones,
-    // and enforces the lower limit itself.
-    'kling-4.0-flash': {
-        textPath: '/text-to-video/kling-4.0-flash',
-        imagePath: '/image-to-video/kling-4.0-flash',
-        audio: 'always',
-        lastFrame: false,
-        maxReferenceImages: 0,
-        maxReferenceImagesWithVideo: 0,
-        maxPromptChars: 26_000,
-    },
-    'kling-4.0': {
-        textPath: '/text-to-video/kling-4.0',
-        imagePath: '/image-to-video/kling-4.0',
-        audio: 'always',
-        lastFrame: true,
-        maxReferenceImages: 0,
-        maxReferenceImagesWithVideo: 0,
-        maxPromptChars: 26_000,
-    },
     'kling-3.0-turbo': {
         textPath: '/text-to-video/kling-3.0-turbo',
         imagePath: '/image-to-video/kling-3.0-turbo',

src/backend/drivers/ai/providers/mistral/chat/MistralAiProvider.test.ts

@@ -83,10 +83,7 @@ let server: PuterServer;
 let recordSpy: MockInstance<MeteringService['utilRecordUsageObject']>;
 
 beforeAll(async () => {
-    // The key puts Mistral on the server's own chat driver.
-    server = await setupTestServer({
-        providers: { mistral: { apiKey: 'test-key' } },
-    });
+    server = await setupTestServer();
 });
 
 afterAll(async () => {
@@ -190,31 +187,11 @@ describe('MistralAIProvider model catalog', () => {
         expect(ids).not.toContain('magistral-medium-latest');
     });
 
-    it.each(['mistral-large-2512', 'mistral-medium-2604'])(
-        'keeps %s, which Mistral serves until it retires the model',
-        async (id) => {
-            const { provider } = makeProvider();
-            expect((await provider.models()).map((model) => model.id)).toContain(
-                id,
-            );
-        },
-    );
-
-    it('moves the floating Large aliases off the deprecated Large 3', async () => {
+    it('keeps mistral-large-2512, which Mistral still serves as active', async () => {
         const { provider } = makeProvider();
-        const models = await provider.models();
-        const aliasesOf = (id: string) =>
-            models.find((model) => model.id === id)?.aliases ?? [];
-        expect(aliasesOf('mistral-large-4')).toEqual(
-            expect.arrayContaining([
-                'mistral-large-latest',
-                'mistral-large',
-                'mistralai/mistral-large-latest',
-            ]),
-        );
-        expect(aliasesOf('mistral-large-2512')).toEqual([
-            'mistralai/mistral-large-2512',
-        ]);
+        const ids = modelLookupNames(await provider.models());
+        expect(ids).toContain('mistral-large-2512');
+        expect(ids).toContain('mistral-large-latest');
     });
 
     it('lists Large 4 with standard pricing and multimodal tool support', async () => {
@@ -227,47 +204,17 @@ describe('MistralAIProvider model catalog', () => {
             modalities: { input: ['text', 'image'], output: ['text'] },
             tool_call: true,
             context: 1_000_000,
-            max_tokens: 262_144,
-            open_weights: true,
+            max_tokens: 1_000_000,
+            open_weights: false,
             costs_currency: 'usd-cents',
             costs: {
                 tokens: 1_000_000,
-                prompt_tokens: 70,
-                cached_tokens: 7,
-                completion_tokens: 210,
+                prompt_tokens: 136,
+                cached_tokens: 14,
+                completion_tokens: 418,
             },
         });
     });
-
-    it.each(['mistral-large-4', 'mistral-large-latest'])(
-        'asks for no more output from %s than Mistral accepts',
-        async (model) => {
-            // Credit for far more output than the context window holds.
-            vi.spyOn(
-                server.services.metering,
-                'getUsageHeadroom',
-            ).mockResolvedValue({ balance: 1e12, held: 0 });
-            completeMock.mockResolvedValueOnce({
-                choices: [
-                    {
-                        message: { content: 'hi', role: 'assistant' },
-                        finishReason: 'stop',
-                    },
-                ],
-                usage: { promptTokens: 1, completionTokens: 1 },
-            });
-
-            await withTestActor(() =>
-                server.drivers.aiChat.complete({
-                    model,
-                    provider: 'mistral',
-                    messages: [{ role: 'user', content: 'hi' }],
-                }),
-            );
-
-            expect(completeMock.mock.calls[0]![0].maxTokens).toBe(262_144);
-        },
-    );
 });
 
 // ── Request shape (Mistral-specific quirks) ─────────────────────────
@@ -592,11 +539,7 @@ describe('MistralAIProvider model resolution', () => {
         ['mistral-large-4-0', 'mistral-large-4'],
         ['mistralai/mistral-large-4', 'mistral-large-4'],
         ['mistralai/mistral-large-4-0', 'mistral-large-4'],
-        ['mistral-large-latest', 'mistral-large-4'],
-        ['mistral-large', 'mistral-large-4'],
-        ['mistralai/mistral-large-latest', 'mistral-large-4'],
-        ['mistral-large-2512', 'mistral-large-2512'],
-        ['mistralai/mistral-large-2512', 'mistral-large-2512'],
+        ['mistral-large-latest', 'mistral-large-2512'],
         ['zai-glm-latest', 'zai-glm-5-3'],
         ['zai-glm-5', 'zai-glm-5-3'],
         ['mistral-code-latest', 'codestral-2508'],
@@ -645,8 +588,8 @@ describe('MistralAIProvider model resolution', () => {
 
 describe('MistralAIProvider.complete non-stream output', () => {
     it.each([
-        ['mistral-large-4', false, 70, 7, 210],
-        ['mistral-large-4', true, 70, 7, 210],
+        ['mistral-large-4', false, 136, 14, 418],
+        ['mistral-large-4', true, 136, 14, 418],
         ['mistral-small-2603', false, 15, 1.5, 60],
         ['mistral-small-2603', true, 15, 1.5, 60],
         ['ministral-8b-2512', false, 15, 1.5, 15],

src/backend/drivers/ai/providers/mistral/chat/models.ts

@@ -25,23 +25,19 @@ export const MISTRAL_MODELS: IChatModel[] = [
         puterId: 'mistralai:mistralai/mistral-large-4',
         id: 'mistral-large-4',
         modalities: { input: ['text', 'image'], output: ['text'] },
-        open_weights: true,
+        open_weights: false,
         tool_call: true,
         release_date: '2026-10-06',
         name: 'Mistral Large 4',
         aliases: [
             'mistral-large-4-0',
-            'mistral-large-latest',
-            'mistral-large',
             'mistralai/mistral-large-4',
             'mistralai/mistral-large-4-0',
-            'mistralai/mistral-large-latest',
         ],
         context: 1_000_000,
-        // Mistral rejects a request for more completion tokens than this.
-        max_tokens: 262_144,
+        max_tokens: 1_000_000,
         description:
-            'State-of-the-art, open-weight, general-purpose multimodal model for reasoning, coding, and agentic workflows.',
+            'Public-preview multimodal model for reasoning, coding, and agentic workflows; open weights are forthcoming.',
         provider: 'mistral',
         costs_currency: 'usd-cents',
         input_cost_key: 'prompt_tokens',
@@ -49,13 +45,12 @@ export const MISTRAL_MODELS: IChatModel[] = [
         serviceSpeedRates: { fast: 1.75 },
         costs: {
             tokens: 1_000_000,
-            prompt_tokens: 70,
-            cached_tokens: 7,
-            completion_tokens: 210,
+            prompt_tokens: 136,
+            cached_tokens: 14,
+            completion_tokens: 418,
         },
     },
     {
-        // Mistral deprecated this model on 2026-10-31 and retires it on 2027-04-30.
         puterId: 'mistralai:mistralai/mistral-medium-2604',
         id: 'mistral-medium-2604',
         modalities: { input: ['text', 'image'], output: ['text'] },
@@ -90,7 +85,6 @@ export const MISTRAL_MODELS: IChatModel[] = [
         },
     },
     {
-        // Mistral deprecated this model on 2026-10-31 and retires it on 2027-04-30.
         puterId: 'mistralai:mistralai/mistral-large-2512',
         id: 'mistral-large-2512',
         modalities: { input: ['text', 'image'], output: ['text'] },
@@ -99,7 +93,12 @@ export const MISTRAL_MODELS: IChatModel[] = [
         knowledge: '2025-12',
         release_date: '2025-12-02',
         name: 'Mistral Large 3',
-        aliases: ['mistralai/mistral-large-2512'],
+        aliases: [
+            'mistral-large-latest',
+            'mistral-large',
+            'mistralai/mistral-large-latest',
+            'mistralai/mistral-large-2512',
+        ],
         context: 262_144,
         max_tokens: 262_144,
         description:

src/backend/drivers/ai/providers/mistral/ocr/MistralOcrProvider.ts

@@ -74,8 +74,7 @@ const toMistralResponseFormat = (
 ): Record<string, unknown> => {
     const format = value as Record<string, unknown> | null;
     const jsonSchema = (format?.jsonSchema ?? format?.json_schema) as
-        | Record<string, unknown>
-        | undefined;
+        Record<string, unknown> | undefined;
     const schema = jsonSchema?.schemaDefinition ?? jsonSchema?.schema;
     if (
         !format ||

src/backend/drivers/ai/providers/openai/speechToText/OpenAISpeechToTextProvider.ts

@@ -155,12 +155,6 @@ export class OpenAISpeechToTextProvider extends SpeechToTextProvider {
                 { legacyCode: 'bad_request' },
             );
         }
-        // The SDK writes a list or an object out as several form fields.
-        if (typeof args.prompt === 'object' && args.prompt !== null) {
-            throw new HttpError(400, 'prompt must be a string', {
-                legacyCode: 'bad_request',
-            });
-        }
         if (args.prompt && !caps.canPrompt) {
             throw new HttpError(
                 400,

src/backend/drivers/ai/shared/fuzzTestUtil.ts

@@ -40,41 +40,9 @@ export const seededRandom = (seed: number): Random => {
     };
 };
 
-/** A case's own sequence, so a case does not depend on the ones before it. */
-export const caseRandom = (seed: number, index: number): Random =>
-    seededRandom(Math.imul(seed ^ index, 0x9e3779b1));
-
 export const pick = <T>(random: Random, values: readonly T[]): T =>
     values[Math.floor(random() * values.length)]!;
 
-/**
- * Picks by rendezvous hashing: a value added to or removed from `values` moves
- * only the cases it wins, so a catalog change leaves every other case alone.
- */
-export const pickStable = <T>(
-    random: Random,
-    values: readonly T[],
-    key: (value: T) => string,
-): T => {
-    const salt = (random() * 0x100000000) >>> 0;
-    let best = values[0]!;
-    let bestScore = -1;
-    for (const value of values) {
-        let hash = salt;
-        for (const char of key(value)) {
-            hash = Math.imul(hash ^ char.charCodeAt(0), 0x01000193);
-        }
-        hash = Math.imul(hash ^ (hash >>> 16), 0x85ebca6b);
-        hash = Math.imul(hash ^ (hash >>> 13), 0xc2b2ae35);
-        const score = (hash ^ (hash >>> 16)) >>> 0;
-        if (score > bestScore) {
-            best = value;
-            bestScore = score;
-        }
-    }
-    return best;
-};
-
 const STRINGS = [
     '',
     ' ',
@@ -308,9 +276,9 @@ export const fuzzCalls = async (options: {
     pinned?: (random: Random) => Record<string, unknown>;
     call: (args: Record<string, unknown>) => Promise<unknown>;
 }): Promise<Findings> => {
+    const random = seededRandom(options.seed);
     const findings = new Findings();
     for (let index = 0; index < options.cases; index++) {
-        const random = caseRandom(options.seed, index);
         const args = {
             ...mutate(
                 pick(random, options.corpus),
@@ -443,9 +411,6 @@ export const upstream = {
     /** Requests `respond` had no reply for. */
     unanswered: new Set<string>(),
 
-    /** How many requests reached `respond`. */
-    asked: 0,
-
     /**
      * Routes global `fetch` here and makes `https.request` throw, so an SDK
      * with its own client fails the suite instead of calling its vendor.
@@ -478,7 +443,6 @@ export const upstream = {
         const text =
             init?.body ??
             (input instanceof Request ? await input.clone().text() : undefined);
-        upstream.asked++;
         const reply = upstream.respond({ url, method, json: parseJson(text) });
         // Stainless SDKs sleep between retries unless told not to.
         const headers: Record<string, string> = { 'x-should-retry': 'false' };
@@ -532,7 +496,6 @@ export const upstream = {
         method?: string;
         data?: unknown;
     }): Promise<{ status: number; data: unknown }> {
-        upstream.asked++;
         const reply = upstream.respond({
             url: new URL(config.url ?? 'http://unknown.test'),
             method: (config.method ?? 'GET').toUpperCase(),

src/backend/drivers/ai/shared/polling.test.ts

@@ -362,7 +362,7 @@ describe('pollUntilSettled', () => {
             const kept: Promise<unknown>[] = [];
             const rejection = rejectionOf(
                 runWithContext({}, () => {
-                    Context.set('keepConcurrencySlot', (work) => kept.push(work));
+                    Context.set('keepDriverSlot', (work) => kept.push(work));
                     return pollUntilSettled<Job>({
                         provider: 'together',
                         providerLabel: 'Together AI',

src/backend/drivers/ai/shared/polling.ts

@@ -184,7 +184,7 @@ export const pollUntilSettled = async <T>(opts: PollOptions<T>): Promise<T> => {
         }
         if (Date.now() - start > windowMs) {
             const late = followLateJob(opts, job, opts.hold.take());
-            Context.get('keepConcurrencySlot')?.(late);
+            Context.get('keepDriverSlot')?.(late);
             throw pollTimeoutError(
                 opts.provider,
                 opts.providerLabel,

src/backend/drivers/util/pipeStreamResult.test.ts

@@ -129,9 +129,9 @@ describe('pipeStreamResult', () => {
         expect(warn).not.toHaveBeenCalled();
     });
 
-    it('adds at most two close listeners to the response', async () => {
-        // The response already carries close listeners from metering,
-        // compression and tracing; Node warns past ten.
+    it('adds no close listener beyond the one pipe adds', async () => {
+        // The response already carries close listeners from middleware and
+        // request tracing; Node warns past ten.
         let added = -1;
         server = createServer((_req, res) => {
             const before = res.listenerCount('close');
@@ -146,8 +146,7 @@ describe('pipeStreamResult', () => {
                 res.on('end', () => resolve());
             }),
         );
-        expect(added).toBeGreaterThan(0);
-        expect(added).toBeLessThanOrEqual(2);
+        expect(added).toBe(1);
     });
 
     it('drops the source, without throwing, when the client left before the result was ready', async () => {

src/backend/drivers/util/pipeStreamResult.ts

@@ -33,9 +33,9 @@ type ResponseSink = NodeJS.WritableStream & {
  * than a clean end, and the failure is logged once instead of surfacing as an
  * unhandled stream error. A client that hangs up stops the source.
  *
- * `pipe` rather than `pipeline`: pipeline puts seven close listeners on the
- * response, and with the ones metering, compression and tracing add, every
- * streamed response tripped Node's ten-listener warning.
+ * `pipe` rather than `pipeline`, and `unpipe` rather than `close`: Node warns
+ * past ten close listeners on the response, and other middleware and request
+ * tracing add their own.
  */
 export function pipeStreamResult(
     stream: Readable,
@@ -52,11 +52,10 @@ export function pipeStreamResult(
         if (!res.destroyed) res.destroy();
         console.warn(`[drivers] ${label} stream failed mid-response:`, err);
     });
-    // A response that closes before the source ended is a client hanging up.
-    res.once('close', () => {
-        if (stream.readableEnded) return;
-        stream.unpipe(res);
-        stream.destroy();
+    // `pipe` unpipes once the response finishes, closes or fails; a source
+    // that hasn't ended by then has a client that hung up.
+    res.once('unpipe', () => {
+        if (!stream.readableEnded) stream.destroy();
     });
     res.once('error', () => stream.destroy());
     stream.pipe(res);

src/backend/server.test.ts

@@ -645,6 +645,59 @@ describe('PuterServer host header validation — permissive modes', () => {
     });
 });
 
+describe('PuterServer behind a trusted proxy', () => {
+    let server: PuterServer;
+    let port: number;
+
+    beforeAll(async () => {
+        port = await allocateEphemeralPort();
+        server = await setupTestServer(
+            {
+                port,
+                domain: 'puter.localhost',
+                origin: `http://puter.localhost:${port}`,
+                api_base_url: `http://api.puter.localhost:${port}`,
+                allow_all_host_values: false,
+                allow_no_host_header: false,
+                custom_domains_enabled: false,
+                static_hosting_domain: 'site.puter.localhost',
+                trust_proxy: 1,
+            } as unknown as IConfig,
+            { listen: true },
+        );
+    });
+
+    afterAll(async () => {
+        await server?.shutdown();
+    });
+
+    it('serves the main domain whatever X-Forwarded-Host names', async () => {
+        const res = await rawRequest(port, '/healthcheck', {
+            host: `puter.localhost:${port}`,
+            'x-forwarded-host': 'nosuchsite.site.puter.localhost',
+        });
+        expect(res.status).toBe(200);
+        expect(res.headers['x-frame-options']).toBe('SAMEORIGIN');
+    });
+
+    it('keeps a hosted-site host on site hosting', async () => {
+        const res = await rawRequest(port, '/healthcheck', {
+            host: `nosuchsite.site.puter.localhost:${port}`,
+            'x-forwarded-host': 'api.puter.localhost',
+        });
+        expect(res.status).toBe(404);
+    });
+
+    it('reads subdomains from Host', async () => {
+        const res = await rawRequest(port, '/healthcheck', {
+            host: `api.puter.localhost:${port}`,
+            'x-forwarded-host': 'puter.localhost',
+        });
+        expect(res.status).toBe(200);
+        expect(res.headers['x-frame-options']).toBeUndefined();
+    });
+});
+
 describe('PuterServer host header validation — custom domains with every host accepted', () => {
     let server: PuterServer;
     let port: number;

src/backend/server.ts

@@ -416,6 +416,17 @@ export class PuterServer {
      *   `#materializeRoute` as those options ship.
      */
     #installGlobalMiddleware() {
+        // -- Forwarded host ------------------------------------------
+        // With `trust proxy` set, express derives `req.hostname` and
+        // `req.subdomains` from X-Forwarded-Host, which any client can send.
+        // Dropping it keeps them on Host, the header host validation checks,
+        // so a request can't pick a different site or subdomain router than
+        // the host it was sent to.
+        this.#app.use((req, _res, next) => {
+            delete req.headers['x-forwarded-host'];
+            next();
+        });
+
         // -- Egress metering -----------------------------------------
         // First, so the byte counter wraps `res.write` before compression
         // does and therefore counts what actually goes out rather than what
@@ -625,7 +636,11 @@ export class PuterServer {
         // `req.deviceFingerprint` (client-supplied, from body/header). Runs
         // AFTER body parsing so the body fingerprint is readable, and before
         // the ALS context so the snapshot carries them.
-        this.#app.use(createFingerprintMiddleware());
+        this.#app.use(
+            createFingerprintMiddleware({
+                tlsFingerprintHeader: this.#config.tlsFingerprintHeader,
+            }),
+        );
 
         // -- Per-request ALS context ---------------------------------
         // Runs AFTER auth probe so `req.actor` is already populated when
@@ -1312,8 +1327,7 @@ export class PuterServer {
         // 2b'. Concurrent in-flight limiting. Same auth-ordering reason
         // (user key + bySubscription resolution needs req.actor); installed
         // after rateLimitGate so a rate-rejection short-circuits before
-        // we acquire a concurrency slot. The slot is released once the response
-        // is done and any work the handler kept it for has settled.
+        // we acquire a concurrency slot. Slot is released on res finish/close.
         if (opts.concurrent) {
             mwChain.push(
                 concurrencyGate(opts.concurrent) as unknown as RequestHandler,

src/backend/services/auth/AuthService.test.ts

@@ -124,6 +124,106 @@ describe('AuthService (integration)', () => {
         return u;
     };
 
+    const makeDebugApp = async (): Promise<string> => {
+        const uid = `app-${uuidv4()}`;
+        await server.clients.db.write(
+            'INSERT INTO `apps` (`uid`, `name`, `title`, `index_url`, `owner_user_id`) VALUES (?, ?, ?, ?, ?)',
+            [uid, `n-${uid}`, `t-${uid}`, `https://${uid}.example/`, 1],
+        );
+        return uid;
+    };
+
+    describe('createDebugAppToken', () => {
+        it('mints an app-under-user token on its own expiring, hidden session', async () => {
+            const user = await makeUser();
+            const appUid = await makeDebugApp();
+            const { token, sessionUuid, expiresAt } =
+                await authService.createDebugAppToken(
+                    user as never,
+                    { uid: appUid },
+                    {
+                        ttlSeconds: 3600,
+                        label: 'debug',
+                        meta: { debug_admin_username: 'op' },
+                    },
+                );
+
+            // Its own dedicated row, not reusing a kind='app' session.
+            const row = (
+                (await server.clients.db.read(
+                    'SELECT `kind`, `app_uid`, `expires_at` FROM `sessions` WHERE `uuid` = ?',
+                    [sessionUuid],
+                )) as Array<Record<string, unknown>>
+            )[0];
+            expect(row?.kind).toBe('debug_app');
+            expect(row?.app_uid).toBe(appUid);
+            expect(Number(row?.expires_at)).toBeGreaterThan(
+                Math.floor(Date.now() / 1000) + 3500,
+            );
+            expect(expiresAt).toBeGreaterThan(Date.now() + 3500 * 1000);
+
+            // Authenticates as the user acting in the app.
+            const { actor } = await authService.authenticate(token);
+            expect(actor?.user.uuid).toBe(user.uuid);
+            expect(actor?.app?.uid).toBe(appUid);
+
+            // Hidden from the user's own session list.
+            const listed = await authService.listSessions({
+                user: { id: user.id },
+            } as Actor);
+            expect(listed.some((s) => s.uuid === sessionUuid)).toBe(false);
+        });
+
+        it('carries an exp, so the token lapses on its own', async () => {
+            const user = await makeUser();
+            const appUid = await makeDebugApp();
+            const { token } = await authService.createDebugAppToken(
+                user as never,
+                { uid: appUid },
+                { ttlSeconds: 1234 },
+            );
+            const claims = server.services.token.verify('auth', token) as {
+                iat: number;
+                exp: number;
+            };
+            expect(claims.exp - claims.iat).toBe(1234);
+        });
+
+        it('records no activity for the user', async () => {
+            const user = await makeUser();
+            const appUid = await makeDebugApp();
+            const { token } = await authService.createDebugAppToken(
+                user as never,
+                { uid: appUid },
+                { ttlSeconds: 600 },
+            );
+            const touch = vi.spyOn(server.stores.session, 'touch');
+            try {
+                const ctx = { ip: '9.9.9.9', userAgent: 'debug' };
+                expect(
+                    (await authService.authenticate(token, ctx)).actor,
+                ).toBeTruthy();
+                expect(touch).not.toHaveBeenCalled();
+            } finally {
+                touch.mockRestore();
+            }
+        });
+
+        it('is evicted by a password reset', async () => {
+            const user = await makeUser();
+            const appUid = await makeDebugApp();
+            const { token } = await authService.createDebugAppToken(
+                user as never,
+                { uid: appUid },
+                { ttlSeconds: 600 },
+            );
+            await authService.revokeInteractiveSessionsForUserId(user.id);
+            const result = await authService.authenticate(token);
+            expect(result.actor).toBeUndefined();
+            expect(result.reauth).toEqual({ reason: 'session_revoked' });
+        });
+    });
+
     describe('authenticateFromToken', () => {
         it('returns null for a malformed/unverifiable token', async () => {
             const actor = await authService.authenticateFromToken('not-a-jwt');

src/backend/services/auth/AuthService.ts

@@ -26,11 +26,13 @@ import {
     type Actor,
 } from '../../core/actor';
 import { HttpError } from '../../core/http/HttpError.js';
+import { normalizeOrigin } from '../../core/http/middleware/originGate.js';
 import { WEB_AND_EXTENSION_PROTOCOLS } from '../../util/validation.js';
 import { awaitLater } from '../../util/concurrency.js';
 import { appGenerationStart } from '../../util/appGeneration.js';
 import {
     ASSET_WINDOW_SECONDS,
+    DEBUG_APP_SESSION_KIND,
     GODMODE_TOKEN_MAX_AGE_SECONDS,
     GODMODE_TOKEN_WINDOW_SECONDS,
     WEB_WINDOW_SECONDS,
@@ -392,6 +394,68 @@ export class AuthService extends PuterService {
         );
     }
 
+    /**
+     * Whether `origin` is one of the deployment's own pages that sign in with a
+     * session of their own (`first_party_session_origins`).
+     */
+    isFirstPartySessionOrigin(origin: unknown): boolean {
+        if (typeof origin !== 'string' || !origin) return false;
+        const wanted = normalizeOrigin(origin);
+        return (this.config.first_party_session_origins ?? []).some(
+            (listed) => normalizeOrigin(listed) === wanted,
+        );
+    }
+
+    /**
+     * Mint the GUI token one of the deployment's own pages signs in with: a
+     * `web` session of its own under the caller's sign-in, so it is listed and
+     * revoked apart from that sign-in and ends with it. Only a sign-in itself
+     * may ask, never a session handed out this way. No session cookie is ever
+     * issued for the row.
+     */
+    async createFirstPartySessionToken(
+        actor: Actor,
+        origin: string,
+        meta: Record<string, unknown> = {},
+    ): Promise<string> {
+        if (
+            !isPlainUserActor(actor) ||
+            !actor.user?.id ||
+            !actor.user.uuid ||
+            actor.session?.kind !== 'web' ||
+            actor.session.parentUid
+        ) {
+            throw new HttpError(
+                403,
+                'Only a signed-in browser session can sign this page in',
+                { legacyCode: 'forbidden' },
+            );
+        }
+        if (!this.isFirstPartySessionOrigin(origin)) {
+            throw new HttpError(403, 'This origin cannot be signed in', {
+                legacyCode: 'forbidden',
+            });
+        }
+
+        const auth_id = this.#authIdFor(actor.user as UserRow);
+        const row = await this.stores.session.getOrCreateFirstPartySession(
+            actor.user.id,
+            {
+                parentSessionUid: actor.session.uid,
+                origin: normalizeOrigin(origin),
+                expiresAt: nowSeconds() + WEB_WINDOW_SECONDS,
+                meta,
+                auth_id,
+            },
+        );
+        if (!row?.uuid) {
+            throw new HttpError(500, 'Could not sign this page in', {
+                legacyCode: 'internal_error',
+            });
+        }
+        return this.createGuiToken(actor.user as UserRow, row.uuid);
+    }
+
     /** Sign a session token for an existing session (upgrade from GUI token). */
     createSessionTokenForSession(user: UserRow, sessionUuid: string): string {
         return this.#signSessionTypeToken(
@@ -787,7 +851,9 @@ export class AuthService extends PuterService {
         const visible = [...live.values()].filter(
             (row) =>
                 row.kind !== 'asset' &&
-                (row.kind !== 'access_token' ||
+                // Operator-created debug sessions are never the user's own.
+                row.kind !== DEBUG_APP_SESSION_KIND &&
+                ((row.kind !== 'access_token' && row.kind !== 'web') ||
                     !row.parent_session_id ||
                     live.has(row.parent_session_id)),
         );
@@ -1075,7 +1141,12 @@ export class AuthService extends PuterService {
         });
         await this.#revokeCascades(
             rows
-                .filter((row) => row.kind === 'web' || row.kind === 'app')
+                .filter(
+                    (row) =>
+                        row.kind === 'web' ||
+                        row.kind === 'app' ||
+                        row.kind === DEBUG_APP_SESSION_KIND,
+                )
                 .map((row) => row.uuid as string),
         );
     }
@@ -1089,16 +1160,24 @@ export class AuthService extends PuterService {
                 legacyCode: 'forbidden',
             });
         }
-        const currentUuid = actor.session?.uid;
-        const rows = await this.stores.session.getByUserId(
+        // The current sign-in is the session, the sign-in it was handed out
+        // under, and the sessions handed out under either.
+        const current = new Set(
+            [actor.session?.uid, actor.session?.parentUid].filter(Boolean),
+        );
+        const rows = (await this.stores.session.getByUserId(
             actor.user.id as number,
             { primary: true },
-        );
+        )) as SessionRow[];
         await this.#revokeCascades(
             rows
                 .filter((row) =>
                     row.kind === 'web'
-                        ? opts.includeCurrent || row.uuid !== currentUuid
+                        ? opts.includeCurrent ||
+                          !(
+                              current.has(row.uuid) ||
+                              current.has(row.parent_session_id ?? '')
+                          )
                         : row.kind === 'app' && opts.includeApps,
                 )
                 .map((row) => row.uuid as string),
@@ -1858,6 +1937,49 @@ export class AuthService extends PuterService {
         });
     }
 
+    /**
+     * An app token acting as `user` in `app` on its own `debug_app` session,
+     * which expires after `ttlSeconds` and never touches the user's real
+     * sessions. The caller authorizes who may ask; `meta` lands on the row.
+     */
+    async createDebugAppToken(
+        user: UserRow,
+        app: { uid: string },
+        opts: {
+            ttlSeconds: number;
+            label?: string | null;
+            meta?: Record<string, unknown>;
+        },
+    ): Promise<{ token: string; sessionUuid: string; expiresAt: number }> {
+        const authId = this.#authIdFor(user);
+        const expiresAtSeconds = nowSeconds() + opts.ttlSeconds;
+        const session = await this.stores.session.create(user.id as number, {
+            kind: DEBUG_APP_SESSION_KIND,
+            app_uid: app.uid,
+            expires_at: expiresAtSeconds,
+            auth_id: authId,
+            label: opts.label ?? null,
+            meta: opts.meta ?? {},
+        });
+        const token = this.services.token.sign(
+            'auth',
+            {
+                type: 'app-under-user',
+                version: '2',
+                user_uid: user.uuid,
+                app_uid: app.uid,
+                session_uid: session.uuid,
+                auth_id: authId,
+            },
+            { expiresIn: opts.ttlSeconds },
+        );
+        return {
+            token,
+            sessionUuid: session.uuid as string,
+            expiresAt: expiresAtSeconds * 1000,
+        };
+    }
+
     /**
      * Mint the full-access token a godmode app runs on. Only the desktop's own
      * web session may ask, and only asking extends it: using the token never
@@ -2702,6 +2824,18 @@ export class AuthService extends PuterService {
         const session: SessionRow | null = rawRow;
 
         if (!session) return { invalid: true };
+        // A session handed to a first-party page lives no longer than the
+        // sign-in it was handed out under.
+        if (session.kind === 'web' && session.parent_session_id) {
+            const parent = (await this.stores.session.getByUuidAny(
+                session.parent_session_id,
+            )) as SessionRow | null;
+            if (!parent) {
+                return { reauth: { reason: 'session_revoked', auth_id } };
+            }
+            const parentEnded = this.#endedSession(parent, auth_id);
+            if (parentEnded) return parentEnded;
+        }
         if (!(await deployed)) return { invalid: true };
 
         void this.#touchSession(session.uuid, user.id, ctx);
@@ -2760,7 +2894,11 @@ export class AuthService extends PuterService {
 
         // An events handler's calls are background work, not the user being
         // active, and its session never expires, so they record no activity.
-        if (decoded.worker_name !== EVENTS_WORKER_SESSION_NAME)
+        // Nor do calls on a debug session, which aren't the user's own.
+        if (
+            decoded.worker_name !== EVENTS_WORKER_SESSION_NAME &&
+            session.kind !== DEBUG_APP_SESSION_KIND
+        )
             void this.#touchSession(session.uuid, user.id, ctx);
 
         const actor = this.#buildAppUnderUserActor(user, app, session);
@@ -2990,7 +3128,13 @@ export class AuthService extends PuterService {
         return makeActor({
             user: this.#actorUserFromRow(user),
             session: session
-                ? { uid: session.uuid, kind: session.kind ?? null }
+                ? {
+                      uid: session.uuid,
+                      kind: session.kind ?? null,
+                      ...(session.kind === 'web' && session.parent_session_id
+                          ? { parentUid: session.parent_session_id }
+                          : {}),
+                  }
                 : null,
         });
     }

src/backend/services/auth/OIDCService.ts

@@ -644,6 +644,7 @@ export class OIDCService extends PuterService {
                     // forwarded chain stays in `audit_metadata.ip_fwd`.
                     signup_ip_forwarded: clientIp,
                     signup_user_agent: req?.headers?.['user-agent'] ?? null,
+                    signup_tls_fingerprint: req?.tlsFingerprint ?? null,
                     signup_origin: req?.headers?.origin,
                     signup_server: this.config.serverId,
                     referrer: referrer ?? null,

src/backend/services/homepage/PuterHomepageService.ts

@@ -399,8 +399,11 @@ export class PuterHomepageService extends PuterService {
             // instead of the socket wire.
             eventsNotifications: notificationsFoldInEnabled(this.config),
             ...meta,
-            // After `meta`, so nothing page-specific can name the opener.
+            // After `meta`, so nothing page-specific can name the opener or
+            // the pages a sign-in popup hands a session.
             ...(popupFrame ? { embed_host_origin: popupFrame.hostOrigin } : {}),
+            first_party_session_origins:
+                this.config.first_party_session_origins ?? [],
         };
         // Only ever this request's; without one the GUI asks for its token as
         // it always has.

src/backend/services/realtime/OpenAIRealtimeRelayService.ts

@@ -1123,15 +1123,13 @@ export class OpenAIRealtimeRelayService extends PuterService {
                     identity?.type === 'app-under-user'
                         ? identity.session
                         : undefined;
-                if (
-                    !(
-                        sessionUids?.has(sessionUid ?? '') ||
-                        sessionUids?.has(parentUid ?? '') ||
-                        (tokenUid &&
-                            identity?.type === 'access-token' &&
-                            identity.token_uid === tokenUid)
-                    )
-                )
+                if (!(
+                    sessionUids?.has(sessionUid ?? '') ||
+                    sessionUids?.has(parentUid ?? '') ||
+                    (tokenUid &&
+                        identity?.type === 'access-token' &&
+                        identity.token_uid === tokenUid)
+                ))
                     return;
                 if (token) {
                     kv.del(realtimeAuthKey(token));

src/backend/services/user/SignupService.ts

@@ -245,6 +245,7 @@ export class SignupService extends PuterService {
             data: input.data,
             ip: clientIp,
             user_agent: req?.headers?.['user-agent'] ?? null,
+            tls_fingerprint: req?.tlsFingerprint ?? null,
             email,
             // The form `email.validate` was given, so the abuse harness can
             // find the verdict that hook cached for this address.
@@ -559,6 +560,7 @@ export class SignupService extends PuterService {
                         // column; the raw chain stays in `audit_metadata`.
                         signup_ip_forwarded: clientIp,
                         signup_user_agent: req.headers?.['user-agent'] ?? null,
+                        signup_tls_fingerprint: req.tlsFingerprint ?? null,
                         signup_origin:
                             (req.headers?.origin as string | null) ?? null,
                         signup_server: this.config.serverId,

src/backend/stores/session/SessionStore.js

@@ -60,6 +60,12 @@ export const APP_WINDOW_SECONDS = 365 * 24 * 60 * 60; // 1y
 export const WORKER_WINDOW_SECONDS = 99 * 365 * 24 * 60 * 60; // 99y (virtually infinite);
 
 export const ASSET_WINDOW_SECONDS = 7 * 24 * 60 * 60; // 7 days
+// An app session created on a user's behalf for debugging: hard expiry, never
+// slid, and hidden from the user's own session list.
+export const DEBUG_APP_SESSION_KIND = 'debug_app';
+// `created_via` of a `web` row handed to one of the deployment's own pages
+// under another sign-in (its `parent_session_id`).
+export const FIRST_PARTY_SESSION_VIA = 'first_party';
 // A godmode app's token lapses at most this long after the desktop last asked
 // for it, so a closed desktop's tokens die on their own. A token past the max
 // age is never renewed again; the next ask replaces it and revokes the old one.
@@ -163,7 +169,8 @@ export class SessionStore extends PuterStore {
      * Interactive sign-ins for a user, newest first, for the account's own
      * activity view. Revoked rows are included -- a session someone opened and
      * closed is exactly what the reader is looking for. Derived kinds (app,
-     * access token, asset, worker) are not sign-ins and stay out.
+     * access token, asset, worker) and sessions handed out under another
+     * sign-in are not sign-ins and stay out.
      *
      * Uncached and keyset-paginated on `id`: the caller merges this with
      * another stream, so it asks for one page at a time rather than the whole
@@ -176,7 +183,7 @@ export class SessionStore extends PuterStore {
     async listSignIns(userId, { limit, beforeId = null }) {
         const rows = await this.clients.db.read(
             'SELECT `id`, `created_at`, `last_ip`, `last_user_agent` FROM `sessions` ' +
-                "WHERE `user_id` = ? AND `kind` = 'web'" +
+                "WHERE `user_id` = ? AND `kind` = 'web' AND `parent_session_id` IS NULL" +
                 (beforeId === null ? '' : ' AND `id` < ?') +
                 ' ORDER BY `id` DESC LIMIT ?',
             beforeId === null ? [userId, limit] : [userId, beforeId, limit],
@@ -490,13 +497,19 @@ export class SessionStore extends PuterStore {
      * sessions costs a few primary round trips rather than a few per session.
      * Returns `{ userId, uuids }` for each user that had rows revoked.
      *
+     * A first-party page's session, revoked under its sign-in, is cascaded from
+     * in turn: it has derived rows of its own.
+     *
      * @param {string[]} rootUuids
      */
     async revokeCascades(rootUuids) {
         const roots = [...new Set(rootUuids.filter(Boolean).map(String))];
+        const queued = new Set(roots);
         const uuidsByUser = new Map();
-        for (let i = 0; i < roots.length; i += REVOKE_CHUNK_SIZE) {
+        // `roots` grows as it is walked, so each pass takes what is there.
+        for (let i = 0; i < roots.length; ) {
             const chunk = roots.slice(i, i + REVOKE_CHUNK_SIZE);
+            i += chunk.length;
             const placeholders = chunk.map(() => '?').join(', ');
             const scope = `(\`uuid\` IN (${placeholders}) OR \`parent_session_id\` IN (${placeholders})) AND \`revoked_at\` IS NULL`;
             const scopeParams = [...chunk, ...chunk];
@@ -528,6 +541,14 @@ export class SessionStore extends PuterStore {
                 const userId = Number(r.user_id);
                 if (!uuidsByUser.has(userId)) uuidsByUser.set(userId, []);
                 uuidsByUser.get(userId).push(String(r.uuid));
+                if (
+                    r.kind === 'web' &&
+                    r.parent_session_id &&
+                    !queued.has(String(r.uuid))
+                ) {
+                    queued.add(String(r.uuid));
+                    roots.push(String(r.uuid));
+                }
             }
         }
         return [...uuidsByUser].map(([userId, uuids]) => ({ userId, uuids }));
@@ -756,6 +777,57 @@ export class SessionStore extends PuterStore {
         return { row, replaced: found ?? null };
     }
 
+    /**
+     * The `web` session one of the deployment's own pages runs on under a
+     * sign-in: the live one for (`parentSessionUid`, `origin`) if there is one,
+     * else a new row. Revoking the parent cascades here. Concurrent first asks
+     * can each insert a row; both stay valid.
+     *
+     * @param {number} userId - User row id.
+     * @param {{
+     *     parentSessionUid: string;
+     *     origin: string;
+     *     expiresAt: number;
+     *     meta?: Record<string, unknown>;
+     *     auth_id?: string;
+     * }} opts
+     *   - `origin` is the page's origin, kept in `meta.first_party_origin`. `meta`
+     *       is the request context, as on any `web` row.
+     */
+    async getOrCreateFirstPartySession(
+        userId,
+        { parentSessionUid, origin, expiresAt, meta = {}, auth_id },
+    ) {
+        if (!userId || !parentSessionUid || !origin) return null;
+
+        const now = nowSeconds();
+        // One per listed origin under a sign-in, so a handful at most.
+        const rows = await this.clients.db.pread(
+            "SELECT * FROM `sessions` WHERE `kind` = 'web' AND `user_id` = ? AND `parent_session_id` = ? AND `created_via` = ? AND `revoked_at` IS NULL ORDER BY `id` DESC LIMIT 20",
+            [userId, parentSessionUid, FIRST_PARTY_SESSION_VIA],
+        );
+        const found = rows
+            .map((row) => this.#normalizeRow(row))
+            .find(
+                (row) =>
+                    row &&
+                    !isExpired(row, now) &&
+                    row.meta?.first_party_origin === origin,
+            );
+        if (found) return found;
+
+        return this.create(userId, {
+            kind: 'web',
+            parent_session_id: parentSessionUid,
+            created_via: FIRST_PARTY_SESSION_VIA,
+            meta: { ...meta, first_party_origin: origin },
+            last_ip: meta.ip ?? null,
+            last_user_agent: meta.user_agent ?? null,
+            expires_at: expiresAt,
+            auth_id,
+        });
+    }
+
     /**
      * Idempotent "give me the app session for this (user, app)" lookup. Returns
      * the existing active app session if one exists, or creates a new one.
@@ -1004,6 +1076,18 @@ export class SessionStore extends PuterStore {
         return result?.affectedRows ?? result?.changes ?? 0;
     }
 
+    /**
+     * Hard-delete debug app sessions that expired before `cutoff` (unix
+     * seconds). Cached copies still read as expired, so need no invalidation.
+     */
+    async deleteExpiredDebugAppSessions(cutoff) {
+        const result = await this.clients.db.write(
+            'DELETE FROM `sessions` WHERE `kind` = ? AND `expires_at` IS NOT NULL AND `expires_at` < ?',
+            [DEBUG_APP_SESSION_KIND, cutoff],
+        );
+        return result?.affectedRows ?? result?.changes ?? 0;
+    }
+
     /**
      * Bump `last_activity` and slide `expires_at` per the row's kind in a
      * single UPDATE. Sliding kinds (web/app/asset) get their `expires_at`

src/backend/stores/session/SessionStore.test.ts

@@ -357,6 +357,28 @@ describe('SessionStore', () => {
         });
     });
 
+    describe('deleteExpiredDebugAppSessions', () => {
+        it('deletes only expired debug_app rows', async () => {
+            const user = await makeUser();
+            const now = Math.floor(Date.now() / 1000);
+            const mk = async (kind: string, expiresAt: number | null) => {
+                const { uuid } = await target.create(user.id, {
+                    kind,
+                    expires_at: expiresAt,
+                });
+                return uuid;
+            };
+            const expired = await mk('debug_app', now - 10);
+            const live = await mk('debug_app', now + 3600);
+            const otherKind = await mk('app', now - 10);
+
+            expect(await target.deleteExpiredDebugAppSessions(now)).toBe(1);
+            expect(await rawRow(expired)).toBeNull();
+            expect(await rawRow(live)).not.toBeNull();
+            expect(await rawRow(otherKind)).not.toBeNull();
+        });
+    });
+
     describe('revokeCascade', () => {
         it('revokes the root session and all child sessions', async () => {
             const user = await makeUser();

src/backend/stores/user/UserStore.ts

@@ -75,6 +75,8 @@ export interface UserRow {
      * "verified a card".
      */
     card_fingerprint?: string | null;
+    /** TLS fingerprint of the request that created the account, if known. */
+    signup_tls_fingerprint?: string | null;
     /**
      * Home region this account's data belongs in - a node id from
      * `IConfig.servers`, not a storage region. Null for accounts that predate
@@ -531,6 +533,7 @@ export class UserStore extends PuterStore {
         signup_ip?: string | null;
         signup_ip_forwarded?: string | null;
         signup_user_agent?: string | null;
+        signup_tls_fingerprint?: string | null;
         signup_origin?: string | null;
         signup_server?: string | null;
         referrer?: string | null;
@@ -557,6 +560,7 @@ export class UserStore extends PuterStore {
              signup_ip,
              signup_ip_forwarded,
              signup_user_agent,
+             signup_tls_fingerprint,
              signup_origin,
              signup_server,
              referrer,
@@ -565,7 +569,7 @@ export class UserStore extends PuterStore {
              phone,
              requires_phone_verification,
              requires_card_verification)
-            VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)${this.clients.db.returningIdClause()}`,
+            VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)${this.clients.db.returningIdClause()}`,
             [
                 fields.username,
                 fields.email,
@@ -585,6 +589,7 @@ export class UserStore extends PuterStore {
                 fields.signup_ip ?? null,
                 fields.signup_ip_forwarded ?? null,
                 fields.signup_user_agent ?? null,
+                fields.signup_tls_fingerprint ?? null,
                 fields.signup_origin ?? null,
                 fields.signup_server ?? null,
                 fields.referrer ?? null,

src/backend/testUtil.ts

@@ -1,5 +1,6 @@
 import bcrypt from 'bcrypt';
 import { mkdtempSync, rmSync } from 'node:fs';
+import { ServerResponse, type IncomingMessage } from 'node:http';
 import { createServer } from 'node:net';
 import type { AddressInfo } from 'node:net';
 import { tmpdir } from 'node:os';
@@ -543,3 +544,56 @@ export const withAiCostFactorHook = async <T>(
         events.off('ai.cost.factor.*', hook);
     }
 };
+
+/**
+ * The most `close` listeners the backend may leave on one response. Request
+ * tracing adds up to five of its own, and Node warns past ten.
+ */
+export const RESPONSE_CLOSE_LISTENER_BUDGET = 5;
+
+/**
+ * Records the most `close` listeners each server response held at once. The
+ * hook is process-wide, so call `restore` when done.
+ */
+export const trackResponseCloseListeners = () => {
+    const peaks = new Map<ServerResponse, number>();
+    const proto = ServerResponse.prototype as unknown as Record<
+        string,
+        (
+            this: ServerResponse,
+            event: string | symbol,
+            listener: unknown,
+        ) => unknown
+    >;
+    const methods = ['on', 'addListener', 'prependListener'];
+    const own = methods.map((m) => Object.getOwnPropertyDescriptor(proto, m));
+    for (const method of methods) {
+        const original = proto[method]!;
+        proto[method] = function (event, listener) {
+            const result = original.call(this, event, listener);
+            if (event === 'close') {
+                const count = this.listenerCount('close');
+                peaks.set(this, Math.max(peaks.get(this) ?? 0, count));
+            }
+            return result;
+        };
+    }
+    return {
+        /** The highest count on any response to a request `match` accepts. */
+        peak: (match: (req: IncomingMessage) => boolean): number =>
+            Math.max(
+                0,
+                ...[...peaks]
+                    .filter(([res]) => match(res.req))
+                    .map(([, count]) => count),
+            ),
+        restore: () => {
+            methods.forEach((method, i) => {
+                const descriptor = own[i];
+                if (descriptor)
+                    Object.defineProperty(proto, method, descriptor);
+                else delete proto[method];
+            });
+        },
+    };
+};

src/backend/types.ts

@@ -269,7 +269,12 @@ export interface IPreludeConfig {
      * an RCS agent provisioned in the Prelude account to actually use RCS.
      */
     preferredChannel?:
-        'sms' | 'rcs' | 'whatsapp' | 'viber' | 'zalo' | 'telegram';
+        | 'sms'
+        | 'rcs'
+        | 'whatsapp'
+        | 'viber'
+        | 'zalo'
+        | 'telegram';
 }
 
 /**
@@ -807,6 +812,12 @@ interface IConfigOptional {
      * forgeable. See https://expressjs.com/en/guide/behind-proxies.html.
      */
     trust_proxy: boolean | number | string | string[];
+    /**
+     * Request header carrying the client's TLS fingerprint (JA3/JA4), set by a
+     * proxy that terminates TLS in front of the server. Only set this when that
+     * proxy overwrites any value the client sends. Unset disables it.
+     */
+    tlsFingerprintHeader?: string;
     /** Don't launch browser when starting. */
     no_browser_launch: boolean;
     /** Disable dev-time frontend webpack watchers. */
@@ -843,6 +854,14 @@ interface IConfigOptional {
      * obtain its token via the AuthMe flow instead.
      */
     allow_gui_origins?: string[];
+    /**
+     * Origins of this deployment's own pages, served apart from the GUI, that
+     * the GUI's sign-in may hand a session of their own instead of an app token
+     * (see `AuthService.createFirstPartySessionToken`). Exact origins only.
+     * Such a page acts as the account, so list only pages this deployment
+     * builds and serves itself.
+     */
+    first_party_session_origins?: string[];
     /** HMAC secret for signed file URLs (/file, /writeFile, /sign). */
     url_signature_secret: string;
     /**
@@ -1405,8 +1424,7 @@ export interface ServerDrainInfo {
 }
 
 export interface WithCostsReporting extends WithLifecycle {
-    getReportedCosts?: () =>
-        // eslint-disable-next-line @typescript-eslint/no-explicit-any
+    getReportedCosts?: () => // eslint-disable-next-line @typescript-eslint/no-explicit-any
         | Promise<Record<string, any>[]>
         // eslint-disable-next-line @typescript-eslint/no-explicit-any
         | Record<string, any>[];

src/backend/util/userDetails.ts

@@ -66,6 +66,7 @@ const SENSITIVE_KEYS: ReadonlySet<string> = new Set([
     'signup_ip',
     'signup_ip_forwarded',
     'signup_user_agent',
+    'signup_tls_fingerprint',
     'signup_origin',
     'signup_server',
     'audit_metadata',

src/docs/src/AI/chat.md

@@ -34,7 +34,7 @@ An object containing the following properties:
 | `stream` | `Boolean` | A boolean indicating whether you want to stream the completion. Defaults to `false`. |
 | `max_tokens` | `Number` | The maximum number of tokens to generate in the completion. By default, the specific model's maximum is used. |
 | `temperature` | `Number` | A number between 0 and 2 indicating the randomness of the completion. Lower values make the output more focused and deterministic, while higher values make it more random. By default, the model's own temperature is used. See [Model-specific features](#model-specific-features). |
-| `reasoning_effort` / `reasoning.effort` | `String` | Controls how much effort reasoning models spend thinking. Supported values: `none`, `minimal`, `low`, `medium`, `high`, `xhigh`, and `max`. Lower values give faster responses with less reasoning. See [Model-specific features](#model-specific-features). |
+| `reasoning_effort` / `reasoning.effort` | `String` | Controls how much effort reasoning models spend thinking. Supported values: `none`, `minimal`, `low`, `medium`, `high`, `xhigh`, and `max`. Lower values give faster responses with less reasoning. OpenAI models, Moonshot's Kimi K3.1 Pro models, Meta's Muse Spark models, `gemini-4-argon`, `gemini-3.1-flash-lite` and `gemini-3-flash-preview` only. See [Model-specific features](#model-specific-features). |
 | `thinking` | `Object \| null` | Extended-thinking control: `{ type: "adaptive" \| "enabled" \| "disabled" \| "between_tools", budgetTokens?, display? }`. Pass `null` to send no thinking configuration at all (the model's own default); leaving it unset falls back to the legacy derivation from `reasoning_effort`. `budgetTokens` only applies to `type: "enabled"`. |
 | `verbosity` / `text.verbosity` | `String` | Controls how long or short responses are. Supported values: `low`, `medium`, and `high`. Lower values give shorter responses. OpenAI models only. |
 | `tools` | `Array` | Function definitions the AI can call. See [Function Calling](#function-calling) for details. |
@@ -437,8 +437,6 @@ The `usageDetails` object on every result carries the same counts under provider
 
 Fast service is available on Claude models with a Fast rate card. Ultrafast falls back to fast. When the account has no Fast quota, the request retries at standard speed. The upstream service tier and inference speed are separate fields.
 
-Pass `speed: "fast"` to request Fast mode, or `speed: "standard"` for standard inference. This option applies to Claude models. An explicit `serviceSpeed` takes precedence.
-
 #### Prompt Caching
 
 Specific to Anthropic models, you can use the cache control feature, allowing you to optimize costs for repeated prompts.
@@ -467,13 +465,11 @@ Fast service uses Priority inference on supported text and image requests. Other
 
 #### `temperature`
 
-Gemini 4 Carbon (`gemini-4-carbon`), Gemini 4 Argon (`gemini-4-argon`) and Gemini 4 Hydrogen (`gemini-4-hydrogen`) ignore `temperature`. The `gemini-3.1-flash-lite` and `gemini-3-flash-preview` models use it.
+`gemini-4-argon` ignores `temperature`. `gemini-3.1-flash-lite` and `gemini-3-flash-preview` use it.
 
 #### `reasoning_effort`
 
-Gemini 4 Carbon and Gemini 4 Argon accept `low`, `medium` and `high`, and use `medium` when you leave it unset. Gemini 4 Hydrogen accepts `low`, `medium`, `high` and `xhigh`, and uses `high` when you leave it unset. The `gemini-3.1-flash-lite` and `gemini-3-flash-preview` models accept `minimal`, `low`, `medium` and `high`, and use Google's default level when you leave it unset. For these five Gemini models, any other value fails with `400 bad_request`, and messages can contain text and images only.
-
-The other Gemini models ignore `reasoning_effort`.
+`gemini-4-argon` accepts `low`, `medium` and `high`, and uses `medium` when you leave it unset. `gemini-3.1-flash-lite` and `gemini-3-flash-preview` accept `minimal`, `low`, `medium` and `high`, and use Google's default level when you leave it unset. For these three Gemini models, any other value fails with `400 bad_request`, and messages can contain text and images only.
 
 #### Function calling
 
@@ -572,7 +568,7 @@ Models with an equivalent Highspeed variant select that variant for fast service
 
 #### `reasoning_effort`
 
-Kimi K3.1 Pro accepts `low`, `high` and `max`, uses `max` when you leave it unset, and fails with `400 bad_request` on any other value. The other Kimi models ignore `reasoning_effort`.
+Kimi K3.1 Pro accepts `low`, `high` and `max`, uses `max` when you leave it unset, and fails with `400 bad_request` on any other value.
 
 #### Prompt Caching
 

src/docs/src/AI/listModels.md

@@ -22,15 +22,12 @@ A string containing the provider you want to list the models for.
 
 Resolves to an array of model objects. Each object always contains `id` and `provider`, and may include fields such as `name`, `aliases`, `context`, `max_tokens`, and a `cost` object (`currency`, `tokens`, `input` and `output` costs in cents). Additional provider-specific capability fields may also be present.
 
-The `puterId` field contains the qualified model name in `<service>:<provider>/<model>` form. Pass this name as the `model` option to [`puter.ai.chat()`](/AI/chat/). The `id` field and aliases also remain accepted model names. When an upstream catalog supplies no provider namespace, the qualified name uses the service name as its namespace.
-
 Example model entry:
 
 ```json
 [
   {
     "id": "claude-opus-4-8",
-    "puterId": "anthropic:anthropic/claude-opus-4-8",
     "provider": "claude",
     "name": "Claude Opus 4.8",
     "aliases": ["claude-opus-4-8-latest"],

src/docs/src/AI/txt2img.md

@@ -233,7 +233,7 @@ The original Replicate entries also support these options:
 | `contrast` | `String` | Leonardo | Contrast preset. |
 | `prompt_enhance` | `Boolean` | Leonardo | Server-side prompt enhancement. |
 
-If you abort a Replicate request, Puter cancels the prediction. A generation that completes before it's cancelled is still metered. Predictions have a 10-minute [upstream deadline](https://replicate.com/docs/topics/predictions/create-a-prediction#prediction-deadlines), after which the call fails with `upstream_timeout` (504). A failed creation request isn't retried. When a model's safety checker rejects the generated image and asks for another run, the prediction is run again up to two more times within the same 10 minutes, unless you set `seed`. Only the run that succeeds is metered. If checking on a running prediction fails, the call fails with `upstream_failed` (502), and the prediction isn't cancelled or metered.
+If you abort a Replicate request, Puter cancels the prediction. A generation that completes before it's cancelled is still metered. Predictions have a 10-minute [upstream deadline](https://replicate.com/docs/topics/predictions/create-a-prediction#prediction-deadlines), after which the call fails with `upstream_timeout` (504). A call starts at most one prediction, since a failed creation request isn't retried. If checking on a running prediction fails, the call fails with `upstream_failed` (502), and the prediction isn't cancelled or metered.
 
 For more details, see the [Replicate API reference](https://replicate.com/docs) and each model's schema page on Replicate.
 

src/docs/src/AI/txt2vid.md

@@ -288,8 +288,6 @@ Available when using a Kling model (provider `'kling-video-generation'`). Every
 
 | Model | Alias | Duration (s) | Resolutions | Audio | Last frame | Reference images | `inputVideo` |
 |-------|-------|--------------|-------------|-------|------------|------------------|--------------|
-| `kling-4.0` | | 3 to 30, default 5 | `720p`, `1080p` | Always | Yes | No | No |
-| `kling-4.0-flash` | | 3 to 20, default 5 | `720p` | Always | No | No | No |
 | `kling-3.0-turbo` (default) | `kling-v3-turbo` | 3 to 15, default 5 | `720p`, `1080p` | Always | No | No | No |
 | `kling-3.0` | `kling-v3` | 3 to 15, default 5 | `720p`, `1080p`, `4k` | Optional | Yes | No | No |
 | `kling-3.0-omni` | `kling-v3-omni` | 3 to 15, default 5 | `720p`, `1080p`, `4k` | Optional | Yes | Up to 7 | Yes |
@@ -301,8 +299,6 @@ The first resolution listed is the default. Prices are per second of output:
 
 | Model | 720p | 1080p | 4K |
 |-------|------|-------|----|
-| `kling-4.0` | $0.157 | $0.2355 | |
-| `kling-4.0-flash` | $0.0942 | | |
 | `kling-3.0-turbo` | $0.112 | $0.14 | |
 | `kling-3.0` | $0.084, $0.126 with audio | $0.112, $0.168 with audio | $0.42 |
 | `kling-3.0-omni` | $0.084, $0.112 with audio, $0.126 with `inputVideo` | $0.112, $0.14 with audio, $0.168 with `inputVideo` | $0.42 |
@@ -315,7 +311,7 @@ A clip is shortened to fit the balance, as described in [Cost and clip length](#
 | Option | Type | Description |
 |--------|------|-------------|
 | `model` | `String` | Video model to use, from the table above (full id or alias) |
-| `prompt` | `String` | Up to 26,000 characters in English or 11,000 in other languages on Kling 4.0 and 4.0 Flash, 3,072 characters on the Kling 3.0 models, 2,500 on the others and on Kling 3.0 Turbo with `input_reference` |
+| `prompt` | `String` | Up to 3,072 characters on the Kling 3.0 models, 2,500 on the others and on Kling 3.0 Turbo with `input_reference` |
 | `seconds` | `Number` | Clip length, from the table above. Any other value falls back to `5`. Ignored with `inputVideo` |
 | `size` | `String` | Resolution tier: `'720p'`, `'1080p'` or `'4k'` (case-insensitive), or `'WIDTHxHEIGHT'`. The `resolution` option is an alias |
 | `width` | `Number` | Optional. Only the ratio of `width` to `height` is used, to pick the aspect ratio: `16:9`, `9:16` or `1:1`. Any other ratio gives `16:9`. A clip from `input_reference` or `inputVideo` takes the image's or video's shape |
@@ -327,8 +323,6 @@ A clip is shortened to fit the balance, as described in [Cost and clip length](#
 
 An option the model lacks, or a combination it doesn't support, is rejected with `bad_request` before any credits are spent. Clips carry no watermark. Kling returns a hosted URL that stays available for 30 days.
 
-On Kling 4.0 and 4.0 Flash, a prompt that mixes English with another language is within the limit when its English characters divided by 26,000 plus its other characters divided by 11,000 come to 1 or less. A prompt over 26,000 characters fails with `bad_request`. Kling rejects a shorter prompt that is over the limit for its language with `upstream_bad_request`. Neither is charged.
-
 Kling 3.0 Omni and Kling O1 edit a video passed as `inputVideo`. The video must be MP4 or QuickTime, up to 200 MiB for a Puter file, and 3 to 15.5 seconds long on Kling 3.0 Omni or 3 to 10 seconds on Kling O1. Kling doesn't take a `data:` URI. It fetches a URL itself, and gets a Puter file as a link that works for 19 minutes, which it fetches when the job starts. A video Kling can't decode fails with `upstream_bad_request` and isn't charged. Up to 4 `reference_images` can go with the video.
 
 For more details, see the [Kling video API reference](https://kling.ai/document-api/api/video/3-0-omni/text-to-video).

src/docs/src/Hosting/create.md

@@ -44,6 +44,19 @@ A `Promise` that will resolve to a [`Subdomain`](/Objects/subdomain/) object whe
 
 If a subdomain with the given name already exists, the promise will be rejected with an error. If the path does not exist, the promise will be rejected with an error.
 
+## Cached and uncached URLs
+
+Each site you create is available at two URLs, `<subdomain>.puter.site` and `<subdomain>.puter.host`.
+
+By default, use `<subdomain>.puter.site`. It's served through a CDN and cached, so it loads fast. Changes to your site's files take a few seconds to show up there.
+
+Use `<subdomain>.puter.host` only in the rare case where you need your files exactly as they are right now and can't wait those few seconds. It isn't served through a CDN or cached, so requests to it can be slower and you're charged more for egress. If you aren't sure which one you need, use `puter.site`.
+
+```
+<subdomain>.puter.site   // cached, use this by default
+<subdomain>.puter.host   // uncached, only for rare cases
+```
+
 ## Examples
 
 ### Create a simple website displaying "Hello world!"

src/docs/src/rate-limits-and-quotas.md

@@ -62,8 +62,6 @@ Shared by chat, image generation, video, TTS, speech and OCR. Each interface and
 | Requests per 10 s   | 200  | 30   | 20        |
 | Concurrent requests | 20   | 3    | 2         |
 
-A chat completion counts as a concurrent request until the model finishes it, even if your app disconnects first, because the model keeps generating and the completion is still billed. That's at most 10 minutes after the disconnect.
-
 | Input                                          | Limit            |
 | ---------------------------------------------- | ---------------- |
 | Text for `txt2speech()`                        | 3,000 characters |
@@ -81,9 +79,7 @@ The standard chat timeout is 10 minutes for a response to start, or to finish if
 
 GPT-6.1 Terra has a 1,050,000-token context window and a maximum output of 128,000 tokens. Requests with more than 272,000 input tokens are charged at twice the input and cache rates and 1.5 times the output rate for the whole request.
 
-Gemini 4 Carbon and Gemini 4 Hydrogen accept text and images within a shared 1,048,576-token window and generate up to 65,536 tokens, including thinking.
-
-A `gemini-4-argon` or `gemini-4-carbon` response that isn't streamed is capped at 32,768 output tokens, and a larger `max_tokens` is lowered to that cap. Streaming uses the model-directory limit.
+A `gemini-4-argon` response that isn't streamed is capped at 32,768 output tokens, and a larger `max_tokens` is lowered to that cap. Streaming uses the model-directory limit.
 
 Text-to-speech calls to ElevenLabs, xAI, Speechify, Kling and Gemini time out after 2 minutes; xAI speech-to-text and voice-changer calls after 10 minutes, upload included. A call that times out before any audio or transcript comes back fails with `504 upstream_timeout`. For Gemini, the 2 minutes cover the whole reply, audio included.
 
@@ -185,7 +181,7 @@ A web search is metered at a flat per-request rate in addition to the tokens it
 | xAI | Up to 5 reference images; more fails with `bad_request`. |
 | Together | Image routes are excluded because they require third-party data sharing; generation fails before any upstream call. |
 | Cloudflare | Output dimensions are clamped per side: FLUX.2 256–1920; Lucid Origin 64–2500; Phoenix 64–2048; SDXL and Inpainting 256–2048. Schnell is fixed at 1024×1024. One reference image on FLUX.2 and Inpainting models. Steps: Schnell 1–8; Lucid Origin 1–40; Phoenix and FLUX.2 Dev 1–50; Klein exactly 4; SDXL and Inpainting 1–20. |
-| Replicate | At most 10 references, or the model's lower limit; FLUX 1.1 Pro accepts one. Each fetched reference is capped at 30 MB. Riverflow accepts at most two fonts. Native options follow each model's schema. For other models with width/height controls, dimensions round to multiples of 8 and clamp to the schema bounds (64–4096 per side if none are given). A call has 10 minutes for its predictions, which are polled every 2 seconds. A prediction that a model's safety checker rejects is run again up to two more times when no `seed` is set. Cancellation cleanup polls for up to 30 seconds, plus an in-flight request. Network timeouts: 90 seconds for creation, 30 seconds for other requests. |
+| Replicate | At most 10 references, or the model's lower limit; FLUX 1.1 Pro accepts one. Each fetched reference is capped at 30 MB. Riverflow accepts at most two fonts. Native options follow each model's schema. For other models with width/height controls, dimensions round to multiples of 8 and clamp to the schema bounds (64–4096 per side if none are given). Predictions expire after 10 minutes and are polled every 2 seconds. Cancellation cleanup polls for up to 30 seconds, plus an in-flight request. Network timeouts: 90 seconds for creation, 30 seconds for other requests. |
 | BytePlus | Pro: 10 references; other Seedream models: 14. |
 | Kling | `kling-v3-omni` and `kling-image-o1`: 10 references; `kling-v3` and `kling-v2-1`: one. Prompt up to 2,500 characters. Each reference is a JPEG or PNG up to 10 MB. Jobs are polled every 2 seconds for up to 5 minutes, then fail with `504 upstream_timeout` and aren't charged. Each call to Kling times out after 60 seconds. |
 | BytePlus explicit output size | Pro: 921,600–4,624,220 total pixels. Lite and 4.5: 3,686,400–16,777,216. 4.0: 921,600–16,777,216. All require integer dimensions and an aspect ratio between 1:16 and 16:1. Sizes under 921,600 total pixels are treated as aspect-ratio hints, not exact sizes. |
@@ -211,7 +207,7 @@ A longer prompt, a larger image or video, or input over the token limit fails wi
 
 The reservation is a conservative estimate based on verified generation usage, not a maximum charge. Google can return more output than the requested 65,536 tokens, and the charge is for the input, video, output and thinking tokens Google reports, even when that's more than the reservation. A reply that reports token usage is charged for it even when the reply failed or has no usable clip.
 
-Kling 4.0 and 4.0 Flash accept a prompt of up to 26,000 characters in English or 11,000 in other languages, and generate 3 to 30 seconds or 3 to 20 seconds respectively. A prompt that mixes English with another language is within the limit when its English characters divided by 26,000 plus its other characters divided by 11,000 come to 1 or less. A prompt over 26,000 characters fails with `bad_request`, and Kling rejects a shorter prompt that is over the limit for its language with `upstream_bad_request`. Neither is charged. Kling 4.0 supports 720p and 1080p, and Flash supports 720p. Other Kling video models take a prompt of up to 3,072 characters (Kling 3.0 models) or 2,500 (the others, and Kling 3.0 Turbo with `input_reference`). Kling 3.0 Omni and Kling O1 take a source video (`inputVideo`) as a URL or as a Puter file of up to 200 MiB, which fails with `413 input_too_large` above that. Kling fetches either one itself. A Puter file is passed as a link that works for 19 minutes. The source must be an MP4 or QuickTime video of 3 to 15.5 seconds (10 on Kling O1). Kling rejects a source outside its limits, or one it can't decode, with `upstream_bad_request`, and the request isn't charged. An edit holds the price of a 16-second clip (10 on Kling O1). Each call to Kling times out after 60 seconds, and a job is polled every 5 seconds.
+Kling video models take a prompt of up to 3,072 characters (Kling 3.0 models) or 2,500 (the others, and Kling 3.0 Turbo with `input_reference`). Kling 3.0 Omni and Kling O1 take a source video (`inputVideo`) as a URL or as a Puter file of up to 200 MiB, which fails with `413 input_too_large` above that. Kling fetches either one itself. A Puter file is passed as a link that works for 19 minutes. The source must be an MP4 or QuickTime video of 3 to 15.5 seconds (10 on Kling O1). Kling rejects a source outside its limits, or one it can't decode, with `upstream_bad_request`, and the request isn't charged. An edit holds the price of a 16-second clip (10 on Kling O1). Each call to Kling times out after 60 seconds, and a job is polled every 5 seconds.
 
 MiniMax video models take a prompt of up to 7,000 characters (H3 models) or 2,000 (Hailuo models), and up to 5 images as frames and references together (2 on H3 Max and Hailuo 02, 1 on Hailuo 2.3 and 2.3 Fast). Hailuo 2.3 Fast requires a first-frame image. Hailuo 02 supports first and last frames together at `768p` or `1080p`, and a first frame alone at `512p`. MiniMax fetches an image URL itself, and a Puter file is sent inline. Each call to MiniMax times out after 60 seconds, and a job is polled every 5 seconds.
 

src/gui/src/UI/UIWindow.js

@@ -24,6 +24,7 @@ import UIAuthLogin from './Auth/UIAuthLogin.js';
 import UIWindowItemProperties from './UIWindowItemProperties.js';
 import new_context_menu_item from '../helpers/newContextMenuItem.js';
 import refresh_item_container from '../helpers/refreshItemContainer.js';
+import { setFolderView } from '../helpers/folderViewPrefs.js';
 import { atPath } from '../helpers/pathElements.js';
 import { resolve_typed_path } from '../helpers/expandHomePath.js';
 import UIWindowAppFeedback from './UIWindowAppFeedback.js';
@@ -3089,9 +3090,6 @@ window.update_window_path = async function (el_window, target_path) {
             $(el_window).removeClass(`window-${ $(el_window).attr('data-uid')}`);
             $(el_window).addClass(`window-${ fsentry.id}`);
             $(el_window).attr('data-uid', fsentry.id);
-            $(el_window).attr('data-sort_by', fsentry.sort_by ?? 'name');
-            $(el_window).attr('data-sort_order', fsentry.sort_order ?? 'asc');
-            $(el_window).attr('data-layout', fsentry.layout ?? 'icons');
             $(el_window_item_container).attr('data-uid', fsentry.id);
             // title - use i18n for system directories
             if ( target_path === window.home_path )
@@ -3132,12 +3130,7 @@ window.update_window_path = async function (el_window, target_path) {
             $(el_window).attr('data-path', target_path);
             $(el_window_navbar_path_input).val(target_path);
             $(el_window_navbar_path_input).attr('data-path', target_path);
-            // update layout
-            window.update_window_layout(el_window, fsentry.layout);
-            // update explore header if in details view
-            if ( fsentry.layout === 'details' ) {
-                window.update_details_layout_sort_visuals(el_window, fsentry.sort_by, fsentry.sort_order);
-            }
+            // The folder's sort and layout are applied by refresh_item_container below.
         }).catch(() => {
             if ( $(el_window).attr('data-path') !== target_path ) {
                 return;
@@ -3489,27 +3482,7 @@ window.set_sort_by = function (item_uid, sort_by, sort_order) {
         sort_order = 'asc';
     }
 
-    $.ajax({
-        url: `${window.api_origin }/set_sort_by`,
-        type: 'POST',
-        data: JSON.stringify({
-            sort_by: sort_by,
-            item_uid: item_uid,
-            sort_order: sort_order,
-        }),
-        async: true,
-        contentType: 'application/json',
-        headers: {
-            'Authorization': `Bearer ${window.auth_token}`,
-        },
-        statusCode: {
-            401: function (xhr) {
-                        window.handle401(xhr);
-                    },
-        },
-        success: function () {
-        },
-    });
+    setFolderView(item_uid, { sort_by, sort_order });
     // update the sort_by & sort_order attr of every matching element
     $(`[data-uid="${item_uid}"]`).attr({
         'data-sort_by': sort_by,
@@ -4987,34 +4960,15 @@ $(document).on('click', '.explore-table-headers-th', function (e) {
 });
 
 window.set_layout = function (item_uid, layout) {
-    $.ajax({
-        url: `${window.api_origin }/set_layout`,
-        type: 'POST',
-        data: JSON.stringify({
-            item_uid: item_uid,
-            layout: layout,
-        }),
-        async: true,
-        contentType: 'application/json',
-        headers: {
-            'Authorization': `Bearer ${window.auth_token}`,
-        },
-        statusCode: {
-            401: function (xhr) {
-                        window.handle401(xhr);
-                    },
-        },
-        success: function () {
-            if ( layout === 'details' ) {
-                let el_window = $(`.window[data-uid="${item_uid}"]`);
-                if ( el_window.length > 0 ) {
-                    let sort_by = el_window.attr('data-sort_by');
-                    let sort_order = el_window.attr('data-sort_order');
-                    window.update_details_layout_sort_visuals(el_window, sort_by, sort_order);
-                }
-            }
-        },
-    });
+    setFolderView(item_uid, { layout });
+    if ( layout === 'details' ) {
+        let el_window = $(`.window[data-uid="${item_uid}"]`);
+        if ( el_window.length > 0 ) {
+            let sort_by = el_window.attr('data-sort_by');
+            let sort_order = el_window.attr('data-sort_order');
+            window.update_details_layout_sort_visuals(el_window, sort_by, sort_order);
+        }
+    }
 };
 
 window.update_details_layout_sort_visuals = function (el_window, sort_by, sort_order) {

src/gui/src/UI/UIWindowManageSessions.js

@@ -47,12 +47,24 @@ const deviceIconSvg = (session) => {
     if ( session.kind === 'worker' ) return ICONS.worker;
     if ( session.kind === 'access_token' ) return ICONS.key;
     if ( session.kind === 'app' ) return ICONS.app;
+    if ( firstPartyHost(session) ) return ICONS.globe;
     const { os } = parseUserAgent(session.last_user_agent);
     if ( os === 'iOS' || os === 'Android' ) return ICONS.phone;
     if ( os ) return ICONS.laptop;
     return ICONS.globe;
 };
 
+// Host of the first-party page a session was handed to under a sign-in, or
+// null for any other session.
+export const firstPartyHost = (session) => {
+    if ( typeof session?.first_party_origin !== 'string' ) return null;
+    try {
+        return new URL(session.first_party_origin).host || null;
+    } catch {
+        return null;
+    }
+};
+
 // Short, human label for the kind pill (falls back to the raw kind).
 const kindBadgeLabel = (kind) => {
     switch ( kind ) {
@@ -303,7 +315,8 @@ const UIWindowManageSessions = async function UIWindowManageSessions (options) {
                 || i18n('ui_session_kind_access_token') || 'Access token';
         }
         if ( session.kind === 'web' ) {
-            return session.label || i18n('ui_session_kind_web') || 'Browser session';
+            return session.label || firstPartyHost(session)
+                || i18n('ui_session_kind_web') || 'Browser session';
         }
         return session.label || session.kind || 'Session';
     };
@@ -320,6 +333,7 @@ const UIWindowManageSessions = async function UIWindowManageSessions (options) {
             sessionTitle(session),
             session.label,
             session.kind,
+            firstPartyHost(session),
             session.last_ip,
             session.app?.title,
             session.app?.name,
@@ -509,6 +523,11 @@ const UIWindowManageSessions = async function UIWindowManageSessions (options) {
             b.classList.add('session-widget-badge', `session-widget-badge-${session.kind}`);
             b.textContent = kindBadgeLabel(session.kind);
             el_badges.appendChild(b);
+        } else if ( firstPartyHost(session) ) {
+            const b = document.createElement('span');
+            b.classList.add('session-widget-badge', 'session-widget-badge-site');
+            b.textContent = i18n('ui_session_kind_site') || 'Site session';
+            el_badges.appendChild(b);
         }
         el_titleline.appendChild(el_badges);
         el_main.appendChild(el_titleline);

src/gui/src/css/style.css

@@ -3988,6 +3988,11 @@ html.dark-mode .usage-table-fade-overlay {
     color: #1d4ed8;
 }
 
+.session-widget-badge-site {
+    background: #dcfce7;
+    color: #15803d;
+}
+
 /* Meta lines */
 .session-widget-meta {
     display: flex;