test(app): restore archive compatibility

.opencode/plugins/tui/discovery-smoke.ts

@@ -1,13 +0,0 @@
-import type { Context } from "../../../packages/plugin/src/tui/context"
-
-export default {
-  id: "test.tui-discovery-smoke",
-  setup(_context: Context) {
-    // context.ui.toast.show({
-    //   title: "TUI plugin discovery works",
-    //   message: "Loaded .opencode/plugins/tui/discovery-smoke.ts",
-    //   variant: "success",
-    //   duration: 30_000,
-    // })
-  },
-}

bun.lock

@@ -147,11 +147,6 @@
         "ws": "8.21.0",
       },
       "devDependencies": {
-        "@ff-labs/fff-bin-darwin-arm64": "0.10.1",
-        "@ff-labs/fff-bin-linux-arm64-gnu": "0.10.1",
-        "@ff-labs/fff-bin-linux-x64-gnu": "0.10.1",
-        "@ff-labs/fff-bin-win32-arm64": "0.10.1",
-        "@ff-labs/fff-bin-win32-x64": "0.10.1",
         "@lydell/node-pty-darwin-arm64": "1.2.0-beta.12",
         "@lydell/node-pty-darwin-x64": "1.2.0-beta.12",
         "@lydell/node-pty-linux-arm64": "1.2.0-beta.12",
@@ -169,11 +164,6 @@
         "@types/bun": "catalog:",
         "@types/semver": "catalog:",
         "@typescript/native-preview": "catalog:",
-        "@yuuang/ffi-rs-darwin-arm64": "1.3.2",
-        "@yuuang/ffi-rs-linux-arm64-gnu": "1.3.2",
-        "@yuuang/ffi-rs-linux-x64-gnu": "1.3.2",
-        "@yuuang/ffi-rs-win32-arm64-msvc": "1.3.2",
-        "@yuuang/ffi-rs-win32-x64-msvc": "1.3.2",
         "vite": "catalog:",
         "vite-plugin-solid": "catalog:",
       },
@@ -374,7 +364,6 @@
         "@effect/platform-node": "catalog:",
         "@effect/sql-sqlite-bun": "catalog:",
         "@ff-labs/fff-bun": "0.10.1",
-        "@ff-labs/fff-node": "0.10.1",
         "@lydell/node-pty": "catalog:",
         "@modelcontextprotocol/sdk": "1.29.0",
         "@opencode-ai/ai": "workspace:*",
@@ -600,7 +589,6 @@
         "zod": "catalog:",
       },
       "devDependencies": {
-        "@opencode-ai/theme": "workspace:*",
         "@opentui/core": "catalog:",
         "@opentui/keymap": "catalog:",
         "@opentui/solid": "catalog:",
@@ -608,22 +596,17 @@
         "@tsconfig/node22": "catalog:",
         "@types/node": "catalog:",
         "@typescript/native-preview": "catalog:",
-        "solid-js": "catalog:",
         "typescript": "catalog:",
       },
       "peerDependencies": {
-        "@opencode-ai/theme": "workspace:*",
         "@opentui/core": ">=0.4.5",
         "@opentui/keymap": ">=0.4.5",
         "@opentui/solid": ">=0.4.5",
-        "solid-js": ">=1.9.0",
       },
       "optionalPeers": [
-        "@opencode-ai/theme",
         "@opentui/core",
         "@opentui/keymap",
         "@opentui/solid",
-        "solid-js",
       ],
     },
     "packages/protocol": {
@@ -876,20 +859,6 @@
         "vite": "catalog:",
       },
     },
-    "packages/theme": {
-      "name": "@opencode-ai/theme",
-      "version": "0.0.0",
-      "dependencies": {
-        "@opentui/core": "catalog:",
-        "effect": "catalog:",
-      },
-      "devDependencies": {
-        "@tsconfig/bun": "catalog:",
-        "@types/bun": "catalog:",
-        "@typescript/native-preview": "catalog:",
-        "typescript": "catalog:",
-      },
-    },
     "packages/tui": {
       "name": "@opencode-ai/tui",
       "version": "1.18.4",
@@ -899,7 +868,6 @@
         "@opencode-ai/plugin": "workspace:*",
         "@opencode-ai/schema": "workspace:*",
         "@opencode-ai/simulation": "workspace:*",
-        "@opencode-ai/theme": "workspace:*",
         "@opencode-ai/ui": "workspace:*",
         "@opencode-ai/util": "workspace:*",
         "@opentui/core": "catalog:",
@@ -1062,7 +1030,6 @@
       },
       "devDependencies": {
         "@astrojs/cloudflare": "14.1.4",
-        "@opencode-ai/theme": "workspace:*",
         "@types/bun": "catalog:",
         "astro": "7.1.3",
         "effect": "catalog:",
@@ -1710,8 +1677,6 @@
 
     "@ff-labs/fff-bun": ["@ff-labs/fff-bun@0.10.1", "", { "optionalDependencies": { "@ff-labs/fff-bin-android-arm64": "0.10.1", "@ff-labs/fff-bin-darwin-arm64": "0.10.1", "@ff-labs/fff-bin-darwin-x64": "0.10.1", "@ff-labs/fff-bin-linux-arm64-gnu": "0.10.1", "@ff-labs/fff-bin-linux-arm64-musl": "0.10.1", "@ff-labs/fff-bin-linux-x64-gnu": "0.10.1", "@ff-labs/fff-bin-linux-x64-musl": "0.10.1", "@ff-labs/fff-bin-win32-arm64": "0.10.1", "@ff-labs/fff-bin-win32-x64": "0.10.1" }, "os": [ "!aix", "!sunos", "!freebsd", "!openbsd", ], "cpu": [ "x64", "arm64", ] }, "sha512-9oUCxypGbf2q3vNfKZ31wdzt5KqjhA9S6TwQaFol/j1lkSHVGvtIa3RvdGHPs1UUuvR/MO7b6pHj054BR9bXPQ=="],
 
-    "@ff-labs/fff-node": ["@ff-labs/fff-node@0.10.1", "", { "dependencies": { "ffi-rs": "^1.0.0" }, "optionalDependencies": { "@ff-labs/fff-bin-android-arm64": "0.10.1", "@ff-labs/fff-bin-darwin-arm64": "0.10.1", "@ff-labs/fff-bin-darwin-x64": "0.10.1", "@ff-labs/fff-bin-linux-arm64-gnu": "0.10.1", "@ff-labs/fff-bin-linux-arm64-musl": "0.10.1", "@ff-labs/fff-bin-linux-x64-gnu": "0.10.1", "@ff-labs/fff-bin-linux-x64-musl": "0.10.1", "@ff-labs/fff-bin-win32-arm64": "0.10.1", "@ff-labs/fff-bin-win32-x64": "0.10.1" }, "os": [ "!aix", "!sunos", "!freebsd", "!openbsd", ], "cpu": [ "x64", "arm64", ] }, "sha512-I2TIWHkey4wpLCncUQymBz/yWrjp/TE91yjrPvSLaPMUMRf3wXwEmelGNnZ5ynxy3cUauYg5b5hsUwhOK4skfA=="],
-
     "@floating-ui/core": ["@floating-ui/core@1.8.0", "", { "dependencies": { "@floating-ui/utils": "^0.2.12" } }, "sha512-0CIZ5itps/8x7BG8dEIhs53BvCUH2PCoogtakwRTut+Arm58sJooJ0AuZhLw2HJYIR5cMLNPBSS728sPho2khQ=="],
 
     "@floating-ui/dom": ["@floating-ui/dom@1.8.0", "", { "dependencies": { "@floating-ui/core": "^1.8.0", "@floating-ui/utils": "^0.2.12" } }, "sha512-yXSrzeHZBTZadLOlfyhCkJHNeLJnHRnRInwdZ40L7ZiaAtrBwoYlsDrX3v5zB1Utk7CLfzcOVnVVWoXEky7Ceg=="],
@@ -2110,8 +2075,6 @@
 
     "@opencode-ai/storybook": ["@opencode-ai/storybook@workspace:packages/storybook"],
 
-    "@opencode-ai/theme": ["@opencode-ai/theme@workspace:packages/theme"],
-
     "@opencode-ai/tui": ["@opencode-ai/tui@workspace:packages/tui"],
 
     "@opencode-ai/ui": ["@opencode-ai/ui@workspace:packages/ui"],
@@ -3288,28 +3251,6 @@
 
     "@xmldom/xmldom": ["@xmldom/xmldom@0.8.13", "", {}, "sha512-KRYzxepc14G/CEpEGc3Yn+JKaAeT63smlDr+vjB8jRfgTBBI9wRj/nkQEO+ucV8p8I9bfKLWp37uHgFrbntPvw=="],
 
-    "@yuuang/ffi-rs-android-arm64": ["@yuuang/ffi-rs-android-arm64@1.3.2", "", { "os": "android", "cpu": "arm64" }, "sha512-eDYLT0kVBkp7e2BwdRDmt6N1rkeDPUHDefk3ZX0/nok+GLsqfy1WBoSL3Yg7HVXN1EyW8OBVc2uK8Zq8HbmaSA=="],
-
-    "@yuuang/ffi-rs-darwin-arm64": ["@yuuang/ffi-rs-darwin-arm64@1.3.2", "", { "os": "darwin", "cpu": "arm64" }, "sha512-kRdgPaOM6TfuC5wHUwstlatk4HNie2lwSLJWQL2LiAUIJ7+96CoiWUNVhwBcFrhdfxhnWenYS6F668CV0vit8Q=="],
-
-    "@yuuang/ffi-rs-darwin-x64": ["@yuuang/ffi-rs-darwin-x64@1.3.2", "", { "os": "darwin", "cpu": "x64" }, "sha512-O3AlVgre8FQcZRJe44Xs7A6iDLumoPXqbw40+eJCa2gyXaXyLPdHoWrS1W9rBCa1QZRRnG7zRulPVFw8C5uo8g=="],
-
-    "@yuuang/ffi-rs-linux-arm-gnueabihf": ["@yuuang/ffi-rs-linux-arm-gnueabihf@1.3.2", "", { "os": "linux", "cpu": "arm" }, "sha512-IXiNdTbIcTCPny5eeElijFWYeKSJjQWSjt9ZyJNdLHYiB1Np+XD6K7wNZS6EOMgMelhW1kQE62T654skGkVDIA=="],
-
-    "@yuuang/ffi-rs-linux-arm64-gnu": ["@yuuang/ffi-rs-linux-arm64-gnu@1.3.2", "", { "os": "linux", "cpu": "arm64" }, "sha512-gWFO6xufUK9lPYUqDvKa6IR243dPqdetgl9Q7HrZWaDu7wLo06QQrosw8QTzndafQnOcBKm6LoLujmGCfTgJOA=="],
-
-    "@yuuang/ffi-rs-linux-arm64-musl": ["@yuuang/ffi-rs-linux-arm64-musl@1.3.2", "", { "os": "linux", "cpu": "arm64" }, "sha512-lejvOSqypPziQH5rzfkDlJ6e92qhWbDutE9ttOO6z5I2k83zoh9iZhZWhaXSU5VqgQpcshRkrbtXb9gy1ft5dA=="],
-
-    "@yuuang/ffi-rs-linux-x64-gnu": ["@yuuang/ffi-rs-linux-x64-gnu@1.3.2", "", { "os": "linux", "cpu": "x64" }, "sha512-s8VCFazaJKmgY2hgMTpWk4TtBY/zy5ovbaGgwyY0FvBD0YvyhcET4IrMsDJpHhFVTPCYfKZ1dN45clD/YiFp6g=="],
-
-    "@yuuang/ffi-rs-linux-x64-musl": ["@yuuang/ffi-rs-linux-x64-musl@1.3.2", "", { "os": "linux", "cpu": "x64" }, "sha512-Ahr5chfKZKWUik20bEZRug+be57LZ2yYrtolyjSRoo7A4ZniBUHBZUNWm6TD6i0CJayqyxWeVk/XiaABD8bY0w=="],
-
-    "@yuuang/ffi-rs-win32-arm64-msvc": ["@yuuang/ffi-rs-win32-arm64-msvc@1.3.2", "", { "os": "win32", "cpu": "arm64" }, "sha512-yhpLcj0qel5VNlpzxPZfNmi7+rEX8444QHjUP6WWLxdRfqPllROu/Cp3OpkBpw3BLdxfcDhWkjWMD5QsJN0Pvg=="],
-
-    "@yuuang/ffi-rs-win32-ia32-msvc": ["@yuuang/ffi-rs-win32-ia32-msvc@1.3.2", "", { "os": "win32", "cpu": [ "x64", "ia32", ] }, "sha512-BFVSbdtg/7mJBw5kQFOPKFiA+SF7z3240HpzHN81Umm4Bp4dWkyx0msYn8+Q7/BBJiLQ4F6bi3Nftk58YA9r9w=="],
-
-    "@yuuang/ffi-rs-win32-x64-msvc": ["@yuuang/ffi-rs-win32-x64-msvc@1.3.2", "", { "os": "win32", "cpu": "x64" }, "sha512-ZL5MJ76n2rjwGo26kCWW7wK6QT/cee00Rx8pfW79pz6vM6jqfhoE7zTnwFiw4aOQUes9+HUc5DeeJ3z+Vb9oLg=="],
-
     "@zip.js/zip.js": ["@zip.js/zip.js@2.7.62", "", {}, "sha512-OaLvZ8j4gCkLn048ypkZu29KX30r8/OfFF2w4Jo5WXFr+J04J+lzJ5TKZBVgFXhlvSkqNFQdfnY1Q8TMTCyBVA=="],
 
     "abbrev": ["abbrev@4.0.0", "", {}, "sha512-a1wflyaL0tHtJSmLSOVybYhy22vRih4eduhhrkcjgrWGnRfrZtovJ2FRjxuTtkkj47O/baf0R86QU5OuYpz8fA=="],
@@ -4074,8 +4015,6 @@
 
     "file-uri-to-path": ["file-uri-to-path@1.0.0", "", {}, "sha512-0Zt+s3L7Vf1biwWZ29aARiVYLx7iMGnEUl9x33fbB/j3jR81u/O2LbqK+Bm1CDSNDKVtJ/YjwY7TUd5SkeLQLw=="],
 
-    "ffi-rs": ["ffi-rs@1.3.2", "", { "optionalDependencies": { "@yuuang/ffi-rs-android-arm64": "1.3.2", "@yuuang/ffi-rs-darwin-arm64": "1.3.2", "@yuuang/ffi-rs-darwin-x64": "1.3.2", "@yuuang/ffi-rs-linux-arm-gnueabihf": "1.3.2", "@yuuang/ffi-rs-linux-arm64-gnu": "1.3.2", "@yuuang/ffi-rs-linux-arm64-musl": "1.3.2", "@yuuang/ffi-rs-linux-x64-gnu": "1.3.2", "@yuuang/ffi-rs-linux-x64-musl": "1.3.2", "@yuuang/ffi-rs-win32-arm64-msvc": "1.3.2", "@yuuang/ffi-rs-win32-ia32-msvc": "1.3.2", "@yuuang/ffi-rs-win32-x64-msvc": "1.3.2" } }, "sha512-4s8dX9VbBw/jd5NOuE3EJRqXaIVdjMyiumeeDzrOhtjQRwp6Bz2za7iksWXTnvTQKV/tTdm1s1w7mObe92zPjQ=="],
-
     "filelist": ["filelist@1.0.6", "", { "dependencies": { "minimatch": "^5.0.1" } }, "sha512-5giy2PkLYY1cP39p17Ech+2xlpTRL9HLspOfEgm0L6CwBXBTgsK5ou0JtzYuepxkaQ/tvhCFIJ5uXo0OrM2DxA=="],
 
     "fill-range": ["fill-range@7.1.1", "", { "dependencies": { "to-regex-range": "^5.0.1" } }, "sha512-YsGpe3WHLK8ZYi4tWDg2Jy3ebRz2rXowDxnld4bkQB00cc/1Zw9AWnC0i9ztDJitivtQvaI9KaLyKrc+hBW0yg=="],

docs/design/v1-v2-database-migration.md

@@ -1,118 +0,0 @@
-# V1 to V2 Database Migration
-
-## Approach
-
-- Use the `dev` branch database schema and migration registry as the V1 baseline.
-- Remove migrations that exist only on the V2 branch.
-- Generate one canonical migration from the `dev` schema to the final V2 schema.
-- Add explicit data operations to that migration where generated DDL is insufficient.
-- Test the migration against a populated database at the exact `dev` schema.
-
-## Preserve
-
-The canonical V1 data remains in its existing tables. In particular, preserve `session`, `message`, and `part` rows.
-
-Preserve `workspace` rows and existing `session.workspace_id` values unchanged. The migration must not clear or rebuild
-workspace relationships.
-
-Keep the `todo` table and its data unchanged. V2 does not currently migrate todos into another representation, and the
-generated migration must not drop the table.
-
-## Truncate
-
-Truncate these pre-launch V2 tables before applying schema changes:
-
-- `event`
-- `event_sequence`
-- `session_message`
-
-These rows are not canonical V1 data. Truncating `event` before adding the required `event.created` column means the
-column needs neither a backfill nor a default. After truncation, rebuild `session_message` from canonical V1 `message`
-and `part` rows rather than retaining its pre-launch V2 contents.
-
-## Message Backfill
-
-Backfill canonical V1 history from `message` and `part` into `session_message`. This is the main data transformation in
-the migration. Preserving the V1 tables alone keeps the data safe but does not make existing history visible through the
-V2 session APIs, which read `session_message`.
-
-Reuse each V1 `message.id` as the corresponding `session_message.id`. Stable IDs keep the migration deterministic and
-avoid rewriting other persisted state that may refer to a message.
-
-Within each session, order V1 messages by `time_created` and then `id`, matching the existing V1 message index. Assign
-contiguous `session_message.seq` values starting at `0`.
-
-Map ordinary V1 messages one-to-one by role. Each ordinary V1 user message becomes one V2 `user` row, and each ordinary
-V1 assistant message becomes one V2 `assistant` row. Fold the source message's ordered V1 parts into that row's V2
-payload.
-
-Handle semantic marker parts before applying the ordinary mapping. In particular, a V1 user message containing a
-`compaction` part and its paired assistant summary represent one compaction operation, not two ordinary messages. Special
-part mappings must be decided explicitly before implementing the backfill.
-
-V1 synthetic content is represented by user text parts with `synthetic: true`, not by a separate message role. A V1 user
-message whose visible text parts are all synthetic should become a V2 `synthetic` message. If a V1 user message mixes
-ordinary and synthetic content, preserve the ordinary content in the V2 `user` row and emit the synthetic content as an
-adjacent V2 `synthetic` row. Ignore text parts marked `ignored`, matching V1 model-history behavior.
-
-Use the V1 compaction user message ID as the ID of the collapsed V2 compaction message. This matches V2's use of the
-admitted compaction input ID and preserves references to the initiating message.
-
-For a completed compaction, create one V2 `compaction` row with `status: "completed"`. Set `reason` from the V1
-compaction part's `auto` flag, join the paired summary assistant's nonempty text parts with blank lines for `summary`, and
-serialize the retained V1 tail beginning at `tail_start_id` for `recent`. Use an empty `recent` value when no tail was
-retained, and use the compaction user message creation time. Do not emit the paired summary assistant as a separate V2
-assistant row.
-
-After rebuilding `session_message`, seed `event_sequence` with one row per migrated session. Set its watermark to that
-session's maximum backfilled `session_message.seq`. This prevents new V2 events from reusing sequence numbers or sorting
-before migrated history. The `event` table remains empty.
-
-## Drop
-
-Drop these pre-launch V2 tables without preserving or transforming their rows:
-
-- `session_input`
-- `session_context_epoch`
-
-Do not transfer `session_input` rows into `session_pending`.
-
-## Create Empty
-
-Let the generated migration create these tables empty:
-
-- `instruction_blob`
-- `instruction_entry`
-- `instruction_state`
-- `session_pending`
-- `kv`
-
-V1 has no canonical data to backfill into these tables. V2 initializes their state as it runs.
-
-## Fork Storage
-
-V1 has no fork-boundary state to backfill. New V2 forks use a required message boundary and persist it in
-`session.fork_boundary`. The durable fork event contains no parent sequence. Its resolved boundary is one of:
-
-- `before`: copy messages before the identified message.
-- `through`: copy messages through the identified message.
-
-Forking an empty session is not supported. `session.fork_seq` and `session.fork_message_id` are not part of the final V2
-schema.
-
-New nullable session columns, including `fork_session_id`, `fork_boundary`, and `time_suspended`, require no explicit
-backfill. Existing rows naturally receive `NULL` when the generated migration adds the columns.
-
-## Verification
-
-The canonical migration test should seed representative V1 sessions, messages, parts, todos, projects, accounts,
-credentials, permissions, shares, and workspaces. After migration, it should verify:
-
-- Preserved rows and encoded values remain unchanged.
-- Todo rows remain available in the unchanged `todo` table.
-- `event` is empty, and stale pre-launch rows are absent from the rebuilt projections.
-- Backfilled `session_message` rows represent the canonical V1 `message` and `part` history.
-- Each migrated session's `event_sequence` watermark matches its maximum backfilled message sequence.
-- Dropped tables no longer exist.
-- New tables exist and are empty.
-- The final schema has no ungenerated changes.

packages/ai/AGENTS.md

@@ -46,7 +46,7 @@ const response = yield * LLMClient.generate(request)
 
 `LLM.request(...)` builds an `LLMRequest`. `LLMClient.generate(...)` reads the executable route carried by `request.model.route`, builds the provider-native body, asks the route's transport for a real `HttpClientRequest.HttpClientRequest`, sends it through `RequestExecutor.Service`, parses the provider stream into common `LLMEvent`s, and finally returns an `LLMResponse`.
 
-Use `LLMClient.stream(request)` when callers want incremental `LLMEvent`s. Use `LLMClient.generate(request)` when callers want those same events collected into an `LLMResponse`.
+Use `LLMClient.stream(request)` when callers want incremental `LLMEvent`s. Use `LLMClient.generate(request)` when callers want those same events collected into an `LLMResponse`. Use `LLMClient.prepare<Body>(request)` to compile a request through the route pipeline without sending it — the optional `Body` type argument narrows `.body` to the route's native shape (e.g. `prepare<OpenAIChatBody>(...)` returns a `PreparedRequestOf<OpenAIChatBody>`). The runtime body is identical; the generic is a type-level assertion.
 
 Filter or narrow `LLMEvent` streams with `LLMEvent.is.*` (camelCase guards, e.g. `events.filter(LLMEvent.is.toolCall)`). The kebab-case `LLMEvent.guards["tool-call"]` form also works but prefer `is.*` in new code.
 
@@ -138,13 +138,13 @@ packages/ai/src/
     ids.ts                  branded IDs, literal types, ProviderMetadata
     options.ts              Generation/Provider/Http options, Limits, Model, cache policy
     messages.ts             content parts, Message, ToolDefinition, LLMRequest
-    events.ts               Usage, individual events, LLMEvent, LLMResponse
+    events.ts               Usage, individual events, LLMEvent, PreparedRequest, LLMResponse
     errors.ts               error reasons, LLMError, ToolFailure
     index.ts                barrel
   llm.ts                    request constructors and convenience helpers
   route/
     index.ts                @opencode-ai/ai/route advanced barrel
-    client.ts               Route.make + LLMClient.stream/generate
+    client.ts               Route.make + LLMClient.prepare/stream/generate
     executor.ts             RequestExecutor service + transport error mapping
     protocol.ts             Protocol type + Protocol.make
     endpoint.ts             Endpoint type + Endpoint.path

packages/ai/README.md

@@ -196,6 +196,7 @@ The hosted result is represented as a provider-executed tool call and tool resul
 - **`LLM.generate` / `LLM.stream`** — re-exported from `LLMClient` for one-import use.
 - **`Message.user(...)` / `Message.assistant(...)` / `Message.tool(...)`** — message constructors from the canonical schema model.
 - **`Model.make(...)` / `ToolCallPart.make(...)` / `ToolResultPart.make(...)` / `ToolDefinition.make(...)`** — model and tool-related constructors from the canonical schema model.
+- **`LLMClient.prepare(request)`** — compile a request through protocol body construction, validation, and HTTP preparation without sending. Useful for inspection and testing.
 - **`LLMEvent.is.*`** — typed guards (`is.textDelta`, `is.toolCall`, `is.finish`, …) for filtering streams.
 - **`Image.generate({...})`** — generate images through a provider-neutral image request and response model.
 - **`ImageClient`** — Effect service and layer for image execution, parallel to `LLMClient`.

packages/ai/example/call-sites.md

@@ -568,7 +568,7 @@ App boundary = explicit durable-config -> typed-provider call
       calling `.model(...)`.
 - [x] Remove request-shaping defaults from `Model`; selected models now carry only
       id, provider, and configured route while defaults live on routes or requests.
-- [x] Rework `LLMClient.stream` / `generate` to read
+- [x] Rework `LLMClient.prepare` / `stream` / `generate` to read
       `request.model.route` directly instead of calling `registeredRoute(...)`.
 - [x] Remove `Route.make(...)` global registration from the normal execution
       path; keep route ids only as diagnostics/provider API labels.

packages/ai/example/tutorial.ts

@@ -50,6 +50,18 @@ const request = LLM.request({
   },
 })
 
+// `http` is intentionally not needed for normal calls. This shows the shape for
+// newly released provider fields before they deserve a typed provider option.
+const rawOverlayExample = LLM.request({
+  model,
+  prompt: "Show the final HTTP overlay shape.",
+  http: {
+    body: { metadata: { example: "tutorial" } },
+    headers: { "x-opencode-tutorial": "1" },
+    query: { debug: "1" },
+  },
+})
+
 // 3. `generate` sends the request and collects the event stream into one
 // response object. `response.text` is the collected text output.
 const generateOnce = Effect.gen(function* () {
@@ -210,15 +222,33 @@ const FakeEcho = {
   }),
 }
 
+// `LLMClient.prepare` is the lower-level inspection hook: it compiles through
+// body conversion, validation, endpoint, auth, and HTTP construction without
+// sending anything over the network.
+const inspectFakeProvider = Effect.gen(function* () {
+  const prepared = yield* LLMClient.prepare(
+    LLM.request({
+      model: FakeEcho.configure().model("tiny-echo"),
+      prompt: "Show me the provider pipeline.",
+    }),
+  )
+
+  console.log("\n== fake provider prepare ==")
+  console.log("route:", prepared.route)
+  console.log("body:", Formatter.formatJson(prepared.body, { space: 2 }))
+})
+
 // Provide the LLM runtime and the HTTP request executor once. Keep one path
-// enabled at a time so the tutorial can demonstrate generate, stream, or
-// tool-loop behavior without spending tokens on every example.
+// enabled at a time so the tutorial can demonstrate generate, prepare, stream,
+// or tool-loop behavior without spending tokens on every example.
 const requestExecutorLayer = RequestExecutor.fetchLayer
 const llmDeps = Layer.mergeAll(requestExecutorLayer, WebSocketExecutor.layer)
 const llmClientLayer = LLMClient.layer.pipe(Layer.provide(llmDeps))
 
 const program = Effect.gen(function* () {
   // yield* generateOnce
+  // yield* inspectFakeProvider
+  // yield* LLMClient.prepare(rawOverlayExample).pipe(Effect.andThen((prepared) => Effect.sync(() => console.log(prepared.body))))
   // yield* streamText
   // yield* generateStructuredObject
   // yield* generateDynamicObject.pipe(Effect.andThen((response) => Effect.sync(() => console.log(response.object))))

packages/ai/package.json

@@ -15,7 +15,6 @@
   ],
   "exports": {
     ".": "./src/index.ts",
-    "./testing": "./src/testing.ts",
     "./*": "./src/*.ts"
   },
   "devDependencies": {

packages/ai/src/llm.ts

@@ -9,28 +9,25 @@ import {
   LLMRequest,
   LLMResponse,
   Message,
-  Model,
   SystemPart,
   ToolChoice,
   ToolDefinition,
   type ContentPart,
-  type ModelProviderOptions,
 } from "./schema"
 import { make as makeTool, toDefinitions, type ToolSchema } from "./tool"
 
 /** Input accepted by `LLM.request`, normalized into the canonical `LLMRequest` class. */
-export type RequestInput<SelectedModel extends Model = Model> = Omit<
+export type RequestInput = Omit<
   ConstructorParameters<typeof LLMRequest>[0],
-  "model" | "system" | "messages" | "tools" | "toolChoice" | "generation" | "http" | "providerOptions"
+  "system" | "messages" | "tools" | "toolChoice" | "generation" | "http" | "providerOptions"
 > & {
-  readonly model: SelectedModel
   readonly system?: string | SystemPart | ReadonlyArray<SystemPart>
   readonly prompt?: string | ContentPart | ReadonlyArray<ContentPart>
   readonly messages?: ReadonlyArray<Message | Message.Input>
   readonly tools?: ReadonlyArray<ToolDefinition.Input>
   readonly toolChoice?: ToolChoice.Input
   readonly generation?: GenerationOptions.Input
-  readonly providerOptions?: NoInfer<ModelProviderOptions<SelectedModel>>
+  readonly providerOptions?: ConstructorParameters<typeof LLMRequest>[0]["providerOptions"]
   readonly http?: HttpOptions.Input
 }
 
@@ -38,7 +35,7 @@ export const generate = LLMClient.generate
 
 export const stream = LLMClient.stream
 
-export const request = <const SelectedModel extends Model>(input: RequestInput<SelectedModel>) => {
+export const request = (input: RequestInput) => {
   const {
     system: requestSystem,
     prompt,
@@ -66,7 +63,7 @@ const GENERATE_OBJECT_TOOL_NAME = "generate_object"
 
 const GENERATE_OBJECT_TOOL_DESCRIPTION = "Return the structured result by calling this tool."
 
-type GenerateObjectBase<SelectedModel extends Model = Model> = Omit<RequestInput<SelectedModel>, "tools" | "toolChoice">
+type GenerateObjectBase = Omit<RequestInput, "tools" | "toolChoice">
 
 export class GenerateObjectResponse<T> {
   constructor(
@@ -83,13 +80,11 @@ export class GenerateObjectResponse<T> {
   }
 }
 
-export interface GenerateObjectOptions<S extends ToolSchema<any>, SelectedModel extends Model = Model>
-  extends GenerateObjectBase<SelectedModel> {
+export interface GenerateObjectOptions<S extends ToolSchema<any>> extends GenerateObjectBase {
   readonly schema: S
 }
 
-export interface GenerateObjectDynamicOptions<SelectedModel extends Model = Model>
-  extends GenerateObjectBase<SelectedModel> {
+export interface GenerateObjectDynamicOptions extends GenerateObjectBase {
   /** Raw JSON Schema object describing the expected output shape. */
   readonly jsonSchema: JsonSchema.JsonSchema
 }
@@ -142,11 +137,11 @@ const runGenerateObject = Effect.fn("LLM.generateObject")(function* (
  * 2. `jsonSchema: JsonSchema.JsonSchema` — `.object` is `unknown`. Use when
  *    the schema is only available at runtime (MCP, plugin manifests). Caller validates.
  */
-export function generateObject<const SelectedModel extends Model, S extends ToolSchema<any>>(
-  options: GenerateObjectOptions<S, SelectedModel>,
+export function generateObject<S extends ToolSchema<any>>(
+  options: GenerateObjectOptions<S>,
 ): Effect.Effect<GenerateObjectResponse<Schema.Schema.Type<S>>, LLMError>
-export function generateObject<const SelectedModel extends Model>(
-  options: GenerateObjectDynamicOptions<SelectedModel>,
+export function generateObject(
+  options: GenerateObjectDynamicOptions,
 ): Effect.Effect<GenerateObjectResponse<unknown>, LLMError>
 export function generateObject(options: GenerateObjectOptions<ToolSchema<any>> | GenerateObjectDynamicOptions) {
   if ("schema" in options) {

packages/ai/src/provider-error.ts

@@ -135,7 +135,7 @@ export function classifyProviderFailure(input: ProviderFailure): LLMError["reaso
       rateLimit: input.rateLimit,
     })
   }
-  if (input.status === 408 || input.status === 409 || (input.status !== undefined && input.status >= 500))
+  if (input.status !== undefined && input.status >= 500)
     return new ProviderInternalReason({
       ...common,
       status: input.status,
@@ -145,6 +145,7 @@ export function classifyProviderFailure(input: ProviderFailure): LLMError["reaso
   if (
     input.status === 400 ||
     input.status === 404 ||
+    input.status === 409 ||
     input.status === 413 ||
     input.status === 422
   )

packages/ai/src/provider-package.ts

@@ -1,4 +1,4 @@
-import type { Model, ProviderOptions } from "./schema"
+import type { Model } from "./schema"
 
 export interface Settings extends Readonly<Record<string, unknown>> {
   readonly headers?: Readonly<Record<string, string>>
@@ -9,11 +9,8 @@ export interface Settings extends Readonly<Record<string, unknown>> {
   }
 }
 
-export interface Definition<
-  ProviderSettings extends Settings = Settings,
-  Options extends ProviderOptions = ProviderOptions,
-> {
-  readonly model: (modelID: string, settings: ProviderSettings) => Model<Options>
+export interface Definition<ProviderSettings extends Settings = Settings> {
+  readonly model: (modelID: string, settings: ProviderSettings) => Model
 }
 
 export * as ProviderPackage from "./provider-package"

packages/ai/src/providers/anthropic-compatible.ts

@@ -47,7 +47,7 @@ export const configure = (input: Config) => {
   })
   return {
     id: ProviderID.make(provider),
-    model: (modelID: string | ModelID) => route.model<AnthropicMessages.ProviderOptionsInput>({ id: modelID }),
+    model: (modelID: string | ModelID) => route.model({ id: modelID }),
     configure,
   }
 }
@@ -57,10 +57,7 @@ export const provider = {
   configure,
 }
 
-export const model: ProviderPackage.Definition<Settings, AnthropicMessages.ProviderOptionsInput>["model"] = (
-  modelID,
-  settings,
-) => {
+export const model: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) => {
   if (settings.apiKey !== undefined && settings.authToken !== undefined)
     throw new Error("Anthropic-compatible apiKey cannot be combined with authToken")
   return configure({

packages/ai/src/providers/anthropic.ts

@@ -52,10 +52,7 @@ export const configure = (input: Config = {}) => {
 }
 
 export const provider = configure()
-export const model: ProviderPackage.Definition<Settings, AnthropicMessages.ProviderOptionsInput>["model"] = (
-  modelID,
-  settings,
-) => {
+export const model: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) => {
   if (settings.apiKey !== undefined && settings.authToken !== undefined)
     throw new Error("Anthropic apiKey cannot be combined with authToken")
   return configure({

packages/ai/src/providers/azure.ts

@@ -99,14 +99,10 @@ export const configure = (input: Config) => {
   const modelDefaults = defaults(input)
 
   const responses = (modelID: string | ModelID) =>
-    configuredResponsesRoute
-      .with(withOpenAIOptions(modelID, modelDefaults))
-      .model<OpenAIProviderOptionsInput>({ id: modelID })
+    configuredResponsesRoute.with(withOpenAIOptions(modelID, modelDefaults)).model({ id: modelID })
 
   const chat = (modelID: string | ModelID) =>
-    configuredChatRoute
-      .with(withOpenAIOptions(modelID, modelDefaults))
-      .model<OpenAIProviderOptionsInput>({ id: modelID })
+    configuredChatRoute.with(withOpenAIOptions(modelID, modelDefaults)).model({ id: modelID })
 
   return {
     id,
@@ -137,12 +133,8 @@ const config = (settings: Settings): Config => {
   throw new Error("Azure requires resourceName or baseURL")
 }
 
-export const responsesModel: ProviderPackage.Definition<Settings, OpenAIProviderOptionsInput>["model"] = (
-  modelID,
-  settings,
-) => configure(config(settings)).responses(modelID)
-export const chatModel: ProviderPackage.Definition<Settings, OpenAIProviderOptionsInput>["model"] = (
-  modelID,
-  settings,
-) => configure(config(settings)).chat(modelID)
+export const responsesModel: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) =>
+  configure(config(settings)).responses(modelID)
+export const chatModel: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) =>
+  configure(config(settings)).chat(modelID)
 export const model = responsesModel

packages/ai/src/providers/cloudflare.ts

@@ -4,7 +4,6 @@ import { Auth } from "../route/auth"
 import { AuthOptions, type AtLeastOne, type ProviderAuthOption } from "../route/auth-options"
 import type { RouteDefaultsInput } from "../route/client"
 import { ProviderID, type ModelID } from "../schema"
-import type { OpenAIProviderOptionsInput } from "./openai-options"
 
 export const aiGatewayID = ProviderID.make("cloudflare-ai-gateway")
 export const workersAIID = ProviderID.make("cloudflare-workers-ai")
@@ -21,11 +20,10 @@ type GatewayURL = AtLeastOne<{
 }
 
 export type AIGatewayOptions = GatewayURL &
-  Omit<RouteDefaultsInput, "providerOptions"> &
+  RouteDefaultsInput &
   ProviderAuthOption<"optional"> & {
     /** Cloudflare AI Gateway authentication token. Sent as `cf-aig-authorization`. */
     readonly gatewayApiKey?: CloudflareSecret
-    readonly providerOptions?: OpenAIProviderOptionsInput
   }
 
 type WorkersAIURL = AtLeastOne<{
@@ -33,11 +31,7 @@ type WorkersAIURL = AtLeastOne<{
   readonly baseURL: string
 }>
 
-export type WorkersAIOptions = WorkersAIURL &
-  Omit<RouteDefaultsInput, "providerOptions"> &
-  ProviderAuthOption<"optional"> & {
-    readonly providerOptions?: OpenAIProviderOptionsInput
-  }
+export type WorkersAIOptions = WorkersAIURL & RouteDefaultsInput & ProviderAuthOption<"optional">
 
 export const aiGatewayBaseURL = (input: GatewayURL) => {
   if (input.baseURL) return input.baseURL
@@ -104,7 +98,7 @@ const configureAIGateway = (options: AIGatewayOptions) => {
   })
   return {
     id: aiGatewayID,
-    model: (modelID: string | ModelID) => route.model<OpenAIProviderOptionsInput>({ id: modelID }),
+    model: (modelID: string | ModelID) => route.model({ id: modelID }),
     configure: configureAIGateway,
   }
 }
@@ -117,7 +111,7 @@ const configureWorkersAI = (options: WorkersAIOptions) => {
   })
   return {
     id: workersAIID,
-    model: (modelID: string | ModelID) => route.model<OpenAIProviderOptionsInput>({ id: modelID }),
+    model: (modelID: string | ModelID) => route.model({ id: modelID }),
     configure: configureWorkersAI,
   }
 }

packages/ai/src/providers/github-copilot.ts

@@ -50,11 +50,9 @@ export const configure = (options: ModelOptions) => {
   const responsesRoute = configuredResponsesRoute(options)
   const chatRoute = configuredChatRoute(options)
   const responses = (modelID: string | ModelID) =>
-    responsesRoute
-      .with(withOpenAIOptions(modelID, defaults(options)))
-      .model<OpenAIProviderOptionsInput>({ id: modelID })
+    responsesRoute.with(withOpenAIOptions(modelID, defaults(options))).model({ id: modelID })
   const chat = (modelID: string | ModelID) =>
-    chatRoute.with(withOpenAIOptions(modelID, defaults(options))).model<OpenAIProviderOptionsInput>({ id: modelID })
+    chatRoute.with(withOpenAIOptions(modelID, defaults(options))).model({ id: modelID })
   return {
     id,
     model: (modelID: string | ModelID) =>

packages/ai/src/providers/google-vertex-chat.ts

@@ -1,9 +1,8 @@
 import type { ProviderPackage } from "../provider-package"
 import { OpenAICompatibleChat } from "../protocols/openai-compatible-chat"
 import type { RouteDefaultsInput } from "../route/client"
-import { ProviderID, type ModelID } from "../schema"
+import { ProviderID, type ModelID, type ProviderOptions } from "../schema"
 import { GoogleVertexShared } from "./google-vertex-shared"
-import type { OpenAIProviderOptionsInput } from "./openai-options"
 
 export const id = ProviderID.make("google-vertex")
 
@@ -12,7 +11,6 @@ export type Config = RouteDefaultsInput &
     readonly baseURL?: string
     readonly location?: string
     readonly project?: string
-    readonly providerOptions?: OpenAIProviderOptionsInput
   }
 
 export interface Settings extends ProviderPackage.Settings {
@@ -21,7 +19,7 @@ export interface Settings extends ProviderPackage.Settings {
   readonly baseURL?: string
   readonly location?: string
   readonly project?: string
-  readonly providerOptions?: OpenAIProviderOptionsInput
+  readonly providerOptions?: ProviderOptions
 }
 
 const route = OpenAICompatibleChat.route.with({
@@ -58,7 +56,7 @@ export const configure = (input: Config = {}) => {
   const route = configuredRoute(input)
   return {
     id,
-    model: (modelID: string | ModelID) => route.model<OpenAIProviderOptionsInput>({ id: modelID }),
+    model: (modelID: string | ModelID) => route.model({ id: modelID }),
     configure,
   }
 }
@@ -68,7 +66,7 @@ export const provider = {
   configure,
 }
 
-export const model: ProviderPackage.Definition<Settings, OpenAIProviderOptionsInput>["model"] = (modelID, settings) => {
+export const model: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) => {
   if (settings.apiKey !== undefined) throw new Error("Google Vertex Chat does not support API keys")
   return configure({
     accessToken: settings.accessToken,

packages/ai/src/providers/google-vertex-messages.ts

@@ -91,7 +91,7 @@ export const configure = (input: Config = {}) => {
   const route = configuredRoute(input)
   return {
     id,
-    model: (modelID: string | ModelID) => route.model<AnthropicMessages.ProviderOptionsInput>({ id: modelID }),
+    model: (modelID: string | ModelID) => route.model({ id: modelID }),
     configure,
   }
 }
@@ -101,10 +101,7 @@ export const provider = {
   configure,
 }
 
-export const model: ProviderPackage.Definition<Settings, AnthropicMessages.ProviderOptionsInput>["model"] = (
-  modelID,
-  settings,
-) => {
+export const model: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) => {
   if (settings.apiKey !== undefined) throw new Error("Google Vertex Messages does not support API keys")
   return configure({
     accessToken: settings.accessToken,

packages/ai/src/providers/google-vertex-responses.ts

@@ -1,9 +1,8 @@
 import type { ProviderPackage } from "../provider-package"
 import { OpenAICompatibleResponses } from "../protocols/openai-compatible-responses"
 import type { RouteDefaultsInput } from "../route/client"
-import { ProviderID, type ModelID } from "../schema"
+import { ProviderID, type ModelID, type ProviderOptions } from "../schema"
 import { GoogleVertexShared } from "./google-vertex-shared"
-import type { OpenResponsesProviderOptionsInput } from "./open-responses-options"
 
 export const id = ProviderID.make("google-vertex")
 
@@ -12,7 +11,6 @@ export type Config = RouteDefaultsInput &
     readonly baseURL?: string
     readonly location?: string
     readonly project?: string
-    readonly providerOptions?: OpenResponsesProviderOptionsInput
   }
 
 export interface Settings extends ProviderPackage.Settings {
@@ -21,7 +19,7 @@ export interface Settings extends ProviderPackage.Settings {
   readonly baseURL?: string
   readonly location?: string
   readonly project?: string
-  readonly providerOptions?: OpenResponsesProviderOptionsInput
+  readonly providerOptions?: ProviderOptions
 }
 
 const route = OpenAICompatibleResponses.route.with({
@@ -60,7 +58,7 @@ export const configure = (input: Config = {}) => {
   const route = configuredRoute(input)
   return {
     id,
-    model: (modelID: string | ModelID) => route.model<OpenResponsesProviderOptionsInput>({ id: modelID }),
+    model: (modelID: string | ModelID) => route.model({ id: modelID }),
     configure,
   }
 }
@@ -70,10 +68,7 @@ export const provider = {
   configure,
 }
 
-export const model: ProviderPackage.Definition<Settings, OpenResponsesProviderOptionsInput>["model"] = (
-  modelID,
-  settings,
-) => {
+export const model: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) => {
   if (settings.apiKey !== undefined) throw new Error("Google Vertex Responses does not support API keys")
   return configure({
     accessToken: settings.accessToken,

packages/ai/src/providers/google-vertex.ts

@@ -77,8 +77,7 @@ const configuredRoute = (input: Config, modelID: string | ModelID) => {
 export const configure = (input: Config = {}) => {
   return {
     id,
-    model: (modelID: string | ModelID) =>
-      configuredRoute(input, modelID).model<Gemini.ProviderOptionsInput>({ id: modelID }),
+    model: (modelID: string | ModelID) => configuredRoute(input, modelID).model({ id: modelID }),
     configure,
   }
 }
@@ -87,10 +86,7 @@ export const provider = {
   id,
   configure,
 }
-export const model: ProviderPackage.Definition<Settings, Gemini.ProviderOptionsInput>["model"] = (
-  modelID,
-  settings,
-) => {
+export const model: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) => {
   if (settings.apiKey !== undefined && settings.accessToken !== undefined)
     throw new Error("Google Vertex apiKey cannot be combined with accessToken or auth")
   return configure({

packages/ai/src/providers/google.ts

@@ -50,14 +50,14 @@ export const configure = (input: Config = {}) => {
     })
   return {
     id,
-    model: (modelID: string | ModelID) => route.model<Gemini.ProviderOptionsInput>({ id: modelID }),
+    model: (modelID: string | ModelID) => route.model({ id: modelID }),
     image,
     configure,
   }
 }
 
 export const provider = configure()
-export const model: ProviderPackage.Definition<Settings, Gemini.ProviderOptionsInput>["model"] = (modelID, settings) =>
+export const model: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) =>
   configure({
     apiKey: settings.apiKey,
     baseURL: settings.baseURL,

packages/ai/src/providers/openai-compatible-responses.ts

@@ -36,7 +36,7 @@ export const configure = (input: Config) => {
   })
   return {
     id: ProviderID.make(provider),
-    model: (modelID: string | ModelID) => route.model<OpenResponsesProviderOptionsInput>({ id: modelID }),
+    model: (modelID: string | ModelID) => route.model({ id: modelID }),
     configure,
   }
 }
@@ -46,10 +46,7 @@ export const provider = {
   configure,
 }
 
-export const model: ProviderPackage.Definition<Settings, OpenResponsesProviderOptionsInput>["model"] = (
-  modelID,
-  settings,
-) =>
+export const model: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) =>
   configure({
     apiKey: settings.apiKey,
     baseURL: settings.baseURL,

packages/ai/src/providers/openai-compatible.ts

@@ -4,15 +4,13 @@ import type { RouteDefaultsInput } from "../route/client"
 import { AuthOptions, type ProviderAuthOption } from "../route/auth-options"
 import type { ProviderPackage } from "../provider-package"
 import { profiles, type OpenAICompatibleProfile } from "./openai-compatible-profile"
-import type { OpenAIProviderOptionsInput } from "./openai-options"
 
 export const id = ProviderID.make("openai-compatible")
 
-type GenericModelOptions = Omit<RouteDefaultsInput, "providerOptions"> &
+type GenericModelOptions = RouteDefaultsInput &
   ProviderAuthOption<"optional"> & {
     readonly provider?: string
     readonly baseURL: string
-    readonly providerOptions?: OpenAIProviderOptionsInput
   }
 
 export interface Settings extends ProviderPackage.Settings {
@@ -21,10 +19,9 @@ export interface Settings extends ProviderPackage.Settings {
   readonly provider?: string
 }
 
-export type FamilyModelOptions = Omit<RouteDefaultsInput, "providerOptions"> &
+export type FamilyModelOptions = RouteDefaultsInput &
   ProviderAuthOption<"optional"> & {
     readonly baseURL?: string
-    readonly providerOptions?: OpenAIProviderOptionsInput
   }
 
 export const routes = [OpenAICompatibleChat.route]
@@ -40,8 +37,7 @@ export const configure = (input: GenericModelOptions) => {
   })
   return {
     id: ProviderID.make(provider),
-    model: (modelID: string | ModelID) =>
-      route.model<OpenAIProviderOptionsInput>({ id: modelID, provider: ProviderID.make(provider) }),
+    model: (modelID: string | ModelID) => route.model({ id: modelID, provider: ProviderID.make(provider) }),
     configure,
   }
 }
@@ -67,7 +63,7 @@ export const provider = {
   configure,
 }
 
-export const model: ProviderPackage.Definition<Settings, OpenAIProviderOptionsInput>["model"] = (modelID, settings) =>
+export const model: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) =>
   configure({
     apiKey: settings.apiKey,
     baseURL: settings.baseURL,

packages/ai/src/providers/openai.ts

@@ -86,15 +86,10 @@ export const configure = (input: Config = {}) => {
   const chatRoute = configuredRoute(OpenAIChat.route, input)
   const modelDefaults = defaults(input)
   const responses = (id: string | ModelID) =>
-    responsesRoute
-      .with(withOpenAIOptions(id, modelDefaults, { textVerbosity: true }))
-      .model<OpenAIProviderOptionsInput>({ id })
+    responsesRoute.with(withOpenAIOptions(id, modelDefaults, { textVerbosity: true })).model({ id })
   const responsesWebSocket = (id: string | ModelID) =>
-    responsesWebSocketRoute
-      .with(withOpenAIOptions(id, modelDefaults, { textVerbosity: true }))
-      .model<OpenAIProviderOptionsInput>({ id })
-  const chat = (id: string | ModelID) =>
-    chatRoute.with(withOpenAIOptions(id, modelDefaults)).model<OpenAIProviderOptionsInput>({ id })
+    responsesWebSocketRoute.with(withOpenAIOptions(id, modelDefaults, { textVerbosity: true })).model({ id })
+  const chat = (id: string | ModelID) => chatRoute.with(withOpenAIOptions(id, modelDefaults)).model({ id })
   const image = (modelID: string | ModelID) =>
     OpenAIImages.model({
       id: modelID,
@@ -137,17 +132,15 @@ const config = (settings: Settings): Config => {
   }
 }
 
-export const model: ProviderPackage.Definition<Settings, OpenAIProviderOptionsInput>["model"] = (modelID, settings) => {
+export const model: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) => {
   const configured = configure(config(settings))
   if (settings.transport === undefined || settings.transport === "http") return configured.responses(modelID)
   if (settings.transport === "websocket") return configured.responsesWebSocket(modelID)
   throw new Error(`Unsupported OpenAI Responses transport: ${String(settings.transport)}`)
 }
 
-export const chatModel: ProviderPackage.Definition<Settings, OpenAIProviderOptionsInput>["model"] = (
-  modelID,
-  settings,
-) => configure(config(settings)).chat(modelID)
+export const chatModel: ProviderPackage.Definition<Settings>["model"] = (modelID, settings) =>
+  configure(config(settings)).chat(modelID)
 export const responses = provider.responses
 export const responsesWebSocket = provider.responsesWebSocket
 export const chat = provider.chat

packages/ai/src/providers/openrouter.ts

@@ -107,7 +107,7 @@ export const configure = (input: ModelOptions = {}) => {
   const route = configuredRoute(input)
   return {
     id,
-    model: (modelID: string | ModelID) => route.model<OpenRouterProviderOptionsInput>({ id: modelID }),
+    model: (modelID: string | ModelID) => route.model({ id: modelID }),
     configure,
   }
 }

packages/ai/src/providers/xai.ts

@@ -5,14 +5,12 @@ import * as OpenAICompatibleProfiles from "./openai-compatible-profile"
 import * as OpenAICompatibleChat from "../protocols/openai-compatible-chat"
 import * as OpenAIResponses from "../protocols/openai-responses"
 import { XAIImages } from "../protocols/xai-images"
-import type { OpenAIProviderOptionsInput } from "./openai-options"
 
 export const id = ProviderID.make("xai")
 
-export type ModelOptions = Omit<RouteDefaultsInput, "providerOptions"> &
+export type ModelOptions = RouteDefaultsInput &
   ProviderAuthOption<"optional"> & {
     readonly baseURL?: string
-    readonly providerOptions?: OpenAIProviderOptionsInput
   }
 
 export type { XAIImageOptions } from "../protocols/xai-images"
@@ -44,8 +42,8 @@ const configuredChatRoute = (input: ModelOptions) => {
 export const configure = (input: ModelOptions = {}) => {
   const responsesRoute = configuredResponsesRoute(input)
   const chatRoute = configuredChatRoute(input)
-  const responses = (modelID: string | ModelID) => responsesRoute.model<OpenAIProviderOptionsInput>({ id: modelID })
-  const chat = (modelID: string | ModelID) => chatRoute.model<OpenAIProviderOptionsInput>({ id: modelID })
+  const responses = (modelID: string | ModelID) => responsesRoute.model({ id: modelID })
+  const chat = (modelID: string | ModelID) => chatRoute.model({ id: modelID })
   const image = (modelID: string | ModelID) =>
     XAIImages.model({
       id: modelID,

packages/ai/src/route/client.ts

@@ -10,7 +10,7 @@ import { WebSocketExecutor } from "./transport"
 import type { Protocol } from "./protocol"
 import { applyCachePolicy } from "../cache-policy"
 import * as ProviderShared from "../protocols/shared"
-import type { LLMError, ProtocolID, ProviderOptions } from "../schema"
+import type { LLMError, PreparedRequestOf, ProtocolID, ProviderOptions } from "../schema"
 import {
   GenerationOptions,
   HttpOptions,
@@ -20,6 +20,7 @@ import {
   ModelLimits,
   LLMError as LLMErrorClass,
   LLMEvent,
+  PreparedRequest,
   ProviderID,
   mergeGenerationOptions,
   mergeHttpOptions,
@@ -45,7 +46,7 @@ export interface Route<Body, Prepared = unknown> {
   readonly defaults: RouteDefaults
   readonly body: RouteBody<Body>
   readonly with: (patch: RoutePatch<Body, Prepared>) => Route<Body, Prepared>
-  readonly model: <Options extends ProviderOptions = ProviderOptions>(input: RouteMappedModelInput) => Model<Options>
+  readonly model: (input: RouteMappedModelInput) => Model
   readonly prepareTransport: (body: Body, request: LLMRequest) => Effect.Effect<Prepared, LLMError>
   readonly streamPrepared: (
     prepared: Prepared,
@@ -92,12 +93,12 @@ export interface RoutePatch<Body, Prepared> extends RouteDefaultsInput {
 
 type RouteMappedModelInput = RouteModelInput | RouteRoutedModelInput
 
-const makeRouteModel = <Options extends ProviderOptions = ProviderOptions>(route: AnyRoute, mapped: RouteMappedModelInput) => {
+const makeRouteModel = (route: AnyRoute, mapped: RouteMappedModelInput) => {
   const provider = route.provider ?? ("provider" in mapped ? mapped.provider : undefined)
   if (!provider) throw new Error(`Route.model(${route.id}) requires a provider`)
   if (!endpointBaseURL(route.endpoint))
     throw new Error(`Route.model(${route.id}) requires an endpoint baseURL — configure it on the route first`)
-  return Model.make<Options>({
+  return Model.make({
     ...mapped,
     provider,
     route,
@@ -141,6 +142,17 @@ export const httpOptions = (input: HttpOptionsInput | undefined) => {
 }
 
 export interface Interface {
+  /**
+   * Compile a request through protocol body construction, validation, and HTTP
+   * preparation without sending it. Returns the prepared request including the
+   * provider-native body.
+   *
+   * Pass a `Body` type argument to statically expose the route's body
+   * shape (e.g. `prepare<OpenAIChatBody>(...)`) — the runtime body is
+   * identical, so this is a type-level assertion the caller makes about which
+   * route the request will resolve to.
+   */
+  readonly prepare: <Body = unknown>(request: LLMRequest) => Effect.Effect<PreparedRequestOf<Body>, LLMError>
   readonly stream: StreamMethod
   readonly generate: GenerateMethod
 }
@@ -284,8 +296,7 @@ function makeFromTransport<Body, Prepared, Frame, Event, State>(
           defaults: mergeRouteDefaults(route.defaults, defaults),
         })
       },
-      model: <Options extends ProviderOptions = ProviderOptions>(input: RouteMappedModelInput) =>
-        makeRouteModel<Options>(route, input),
+      model: (input) => makeRouteModel(route, input),
       prepareTransport: (body, request) =>
         routeInput.transport.prepare({
           body,
@@ -359,6 +370,9 @@ export function make<Body, Prepared, Frame, Event, State>(
   })
 }
 
+// `compile` is the important boundary: it turns a common `LLMRequest` into a
+// validated provider body plus transport-private prepared data, but does not
+// execute transport.
 const compile = Effect.fn("LLM.compile")(function* (request: LLMRequest) {
   const resolved = applyCachePolicy(resolveRequestOptions(request))
   const route = resolved.model.route
@@ -376,17 +390,17 @@ const compile = Effect.fn("LLM.compile")(function* (request: LLMRequest) {
   }
 })
 
-/** @internal Test-only projection of the execution compiler; not exported from package barrels. */
-export const compileRequest = Effect.fn("LLM.compileRequest")(function* (request: LLMRequest) {
+const prepareWith = Effect.fn("LLMClient.prepare")(function* (request: LLMRequest) {
   const compiled = yield* compile(request)
-  return {
+
+  return new PreparedRequest({
     id: compiled.request.id ?? "request",
     route: compiled.route.id,
     protocol: compiled.route.protocol,
     model: compiled.request.model,
     body: compiled.body,
     metadata: { transport: compiled.route.transport.id },
-  }
+  })
 })
 
 const streamRequestWith = (runtime: TransportRuntime) => (request: LLMRequest) =>
@@ -408,6 +422,9 @@ const generateWith = (stream: Interface["stream"]) =>
     )
   })
 
+export const prepare = <Body = unknown>(request: LLMRequest) =>
+  prepareWith(request) as Effect.Effect<PreparedRequestOf<Body>, LLMError>
+
 export function stream(request: LLMRequest): Stream.Stream<LLMEvent, LLMError> {
   return Stream.unwrap(
     Effect.gen(function* () {
@@ -436,7 +453,7 @@ export const layer: Layer.Layer<Service, never, RequestExecutor.Service> = Layer
       http: yield* RequestExecutor.Service,
       webSocket: Option.getOrUndefined(yield* Effect.serviceOption(WebSocketExecutor.Service)),
     })
-    return Service.of({ stream, generate: generateWith(stream) })
+    return Service.of({ prepare: prepareWith as Interface["prepare"], stream, generate: generateWith(stream) })
   }),
 )
 
@@ -445,6 +462,7 @@ export const Route = { make } as const
 export const LLMClient = {
   Service,
   layer,
+  prepare,
   stream,
   generate,
 } as const

packages/ai/src/schema/events.ts

@@ -1,5 +1,6 @@
 import { Schema } from "effect"
-import { ContentBlockID, FinishReason, ProviderMetadata, ToolCallID } from "./ids"
+import { ContentBlockID, FinishReason, ProtocolID, ProviderMetadata, RouteID, ToolCallID } from "./ids"
+import { ModelSchema } from "./options"
 import { Message, ToolCallPart, ToolOutput, ToolResultPart, ToolResultValue, type ContentPart } from "./messages"
 import { ProviderFailureClassification } from "./errors"
 
@@ -313,6 +314,29 @@ export const LLMEvent = Object.assign(llmEventTagged, {
 })
 export type LLMEvent = Schema.Schema.Type<typeof llmEventTagged>
 
+export class PreparedRequest extends Schema.Class<PreparedRequest>("LLM.PreparedRequest")({
+  id: Schema.String,
+  route: RouteID,
+  protocol: ProtocolID,
+  model: ModelSchema,
+  body: Schema.Unknown,
+  metadata: Schema.optional(Schema.Record(Schema.String, Schema.Unknown)),
+}) {}
+
+/**
+ * A `PreparedRequest` whose `body` is typed as `Body`. Use with the generic
+ * on `LLMClient.prepare<Body>(...)` when the caller knows which route their
+ * request will resolve to and wants its native shape statically exposed
+ * (debug UIs, request previews, plan rendering).
+ *
+ * The runtime body is identical — the route still emits `body: unknown` — so
+ * this is a type-level assertion the caller makes about what they expect to
+ * find. The prepare runtime does not validate the assertion.
+ */
+export type PreparedRequestOf<Body> = Omit<PreparedRequest, "body"> & {
+  readonly body: Body
+}
+
 const responseText = (events: ReadonlyArray<LLMEvent>) =>
   events
     .filter(LLMEvent.is.textDelta)

packages/ai/src/schema/options.ts

@@ -178,8 +178,7 @@ export namespace ModelCompatibility {
   export const make = (input: Input) => (input instanceof ModelCompatibility ? input : new ModelCompatibility(input))
 }
 
-export class Model<Options extends ProviderOptions = ProviderOptions> {
-  declare protected readonly _ProviderOptions: Options
+export class Model {
   readonly id: ModelID
   readonly provider: ProviderID
   readonly route: AnyRoute
@@ -194,8 +193,8 @@ export class Model<Options extends ProviderOptions = ProviderOptions> {
     this.compatibility = input.compatibility
   }
 
-  static make<Options extends ProviderOptions = ProviderOptions>(input: Model.Input) {
-    return new Model<Options>({
+  static make(input: Model.Input) {
+    return new Model({
       id: ModelID.make(input.id),
       provider: ProviderID.make(input.provider),
       route: input.route,
@@ -204,7 +203,7 @@ export class Model<Options extends ProviderOptions = ProviderOptions> {
     })
   }
 
-  static input<Options extends ProviderOptions>(model: Model<Options>): Model.ConstructorInput {
+  static input(model: Model): Model.ConstructorInput {
     return {
       id: model.id,
       provider: model.provider,
@@ -214,9 +213,9 @@ export class Model<Options extends ProviderOptions = ProviderOptions> {
     }
   }
 
-  static update<Options extends ProviderOptions>(model: Model<Options>, patch: Partial<Model.Input>) {
+  static update(model: Model, patch: Partial<Model.Input>) {
     if (Object.keys(patch).length === 0) return model
-    return Model.make<Options>({
+    return Model.make({
       ...Model.input(model),
       ...patch,
     })
@@ -242,8 +241,6 @@ export namespace Model {
 
 export type ModelInput = Model.Input
 
-export type ModelProviderOptions<SelectedModel> = SelectedModel extends Model<infer Options> ? Options : never
-
 export const ModelSchema = Schema.declare((value): value is Model => value instanceof Model, { expected: "LLM.Model" })
 
 export class CacheHint extends Schema.Class<CacheHint>("LLM.CacheHint")({

packages/ai/src/testing.ts

@@ -1,156 +0,0 @@
-export * as TestLLM from "./testing"
-
-import { LLMClient, type Interface as LLMClientShape } from "./route/client"
-import {
-  LLMEvent,
-  LLMResponse,
-  type FinishReasonDetails,
-  type LLMError,
-  type LLMRequest,
-  type UsageInput,
-} from "./schema"
-import { Context, Deferred, Effect, Latch, Layer, Queue, Scope, Stream } from "effect"
-
-export type Response = readonly LLMEvent[] | Stream.Stream<LLMEvent, LLMError>
-
-export type Gate = Readonly<{ started: Effect.Effect<void>; release: Effect.Effect<void> }>
-
-export interface Interface {
-  readonly requests: LLMRequest[]
-  readonly push: (...responses: readonly Response[]) => Effect.Effect<void>
-  readonly always: (response: Response) => Effect.Effect<void>
-  readonly wait: (count: number) => Effect.Effect<void>
-  readonly gate: Effect.Effect<Gate, never, Scope.Scope>
-  readonly client: LLMClientShape
-}
-
-export interface LayerOptions {
-  readonly transformRequest?: (request: LLMRequest) => LLMRequest
-  /** Used after the one-shot response queue is exhausted. Omit to defect on unexpected requests. */
-  readonly fallback?: Response
-}
-
-export class Service extends Context.Service<Service, Interface>()("@opencode/ai/TestLLM") {}
-
-export const complete = (
-  options: { readonly reason: FinishReasonDetails; readonly usage?: UsageInput },
-  ...events: readonly LLMEvent[]
-) => [
-  LLMEvent.stepStart({ index: 0 }),
-  ...events,
-  LLMEvent.stepFinish({ index: 0, reason: options.reason, usage: options.usage }),
-  LLMEvent.finish({ reason: options.reason }),
-]
-
-export const stop = (...events: readonly LLMEvent[]) => complete({ reason: { normalized: "stop" } }, ...events)
-
-export const toolCalls = (...events: readonly LLMEvent[]) =>
-  complete({ reason: { normalized: "tool-calls" } }, ...events)
-
-const textEvents = (value: string, id: string) => [
-  LLMEvent.textStart({ id }),
-  LLMEvent.textDelta({ id, text: value }),
-  LLMEvent.textEnd({ id }),
-]
-
-export const text = (value: string, id: string) => stop(...textEvents(value, id))
-
-export const textWithUsage = (value: string, id: string, inputTokens: number) =>
-  complete(
-    { reason: { normalized: "stop" }, usage: { inputTokens, nonCachedInputTokens: inputTokens } },
-    ...textEvents(value, id),
-  )
-
-export const tool = (id: string, name: string, input: unknown) => toolCalls(LLMEvent.toolCall({ id, name, input }))
-
-export const failAfter = (error: LLMError, ...events: readonly LLMEvent[]) =>
-  Stream.fromIterable(events).pipe(Stream.concat(Stream.fail(error)))
-
-export const hangAfter = (...events: readonly LLMEvent[]) => Stream.concat(Stream.fromIterable(events), Stream.never)
-
-const toStream = (response: Response) => (Stream.isStream(response) ? response : Stream.fromIterable(response))
-
-export const layer = (options: LayerOptions = {}) =>
-  Layer.effect(
-    Service,
-    Effect.gen(function* () {
-      const requests: LLMRequest[] = []
-      const responses: Response[] = []
-      let started = Deferred.makeUnsafe<void>()
-      let fallback = options.fallback
-      let activeGate: { readonly started: Queue.Queue<void>; readonly release: Latch.Latch } | undefined
-      const wait = (count: number): Effect.Effect<void> =>
-        Effect.suspend(() =>
-          requests.length >= count ? Effect.void : Deferred.await(started).pipe(Effect.andThen(wait(count))),
-        )
-
-      const stream = ((request: LLMRequest) => {
-        requests.push(options.transformRequest?.(request) ?? request)
-        const waiting = started
-        started = Deferred.makeUnsafe()
-        Deferred.doneUnsafe(waiting, Effect.void)
-        const response = responses.shift() ?? fallback
-        if (!response) return Stream.die(new Error(`TestLLM has no response for request ${requests.length}`))
-        const streamed = toStream(response)
-        const gate = activeGate
-        if (!gate) return streamed
-        return Stream.unwrap(
-          Queue.offer(gate.started, undefined).pipe(Effect.andThen(gate.release.await), Effect.as(streamed)),
-        )
-      }) as LLMClientShape["stream"]
-      const client = LLMClient.Service.of({
-        stream,
-        generate: (request) =>
-          stream(request).pipe(
-            Stream.runFold(LLMResponse.empty, LLMResponse.reduce),
-            Effect.flatMap((state) => {
-              const response = LLMResponse.complete(state)
-              if (response) return Effect.succeed(response)
-              return Effect.die("TestLLM response ended without a terminal finish event")
-            }),
-          ),
-      })
-
-      return Service.of({
-        requests,
-        push: (...input) =>
-          Effect.sync(() => {
-            responses.push(...input)
-          }),
-        always: (response) =>
-          Effect.sync(() => {
-            fallback = response
-          }),
-        wait,
-        gate: Effect.gen(function* () {
-          const gate = {
-            started: yield* Effect.acquireRelease(Queue.unbounded<void>(), Queue.shutdown),
-            release: yield* Latch.make(),
-          }
-          activeGate = gate
-          const release = Effect.sync(() => {
-            if (activeGate === gate) activeGate = undefined
-          }).pipe(Effect.andThen(gate.release.open), Effect.asVoid)
-          yield* Effect.addFinalizer(() => release)
-          return {
-            started: Queue.take(gate.started),
-            release,
-          }
-        }),
-        client,
-      })
-    }),
-  )
-
-export const clientLayer = Layer.effect(
-  LLMClient.Service,
-  Effect.map(Service, (service) => service.client),
-)
-
-export const push = (...responses: readonly Response[]) => Service.use((service) => service.push(...responses))
-
-export const always = (response: Response) => Service.use((service) => service.always(response))
-
-export const wait = (count: number) => Service.use((service) => service.wait(count))
-
-export const gate = Service.use((service) => service.gate)

packages/ai/test/adapter.test.ts

@@ -2,7 +2,6 @@ import { describe, expect } from "bun:test"
 import { Effect, Schema, Stream } from "effect"
 import { LLM, LLMRequest, LLMResponse } from "../src"
 import { Route, Endpoint, LLMClient, Protocol, type FramingDef } from "../src/route"
-import { compileRequest } from "../src/route/client"
 import { Model } from "../src/schema"
 import { testEffect } from "./lib/effect"
 import { dynamicResponse } from "./lib/http"
@@ -140,7 +139,8 @@ describe("llm route", () => {
 
   it.effect("selects routes by model route value", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const llm = yield* LLMClient.Service
+      const prepared = yield* llm.prepare(
         LLMRequest.update(request, { model: updateModel(request.model, { route: configuredGemini }) }),
       )
 
@@ -173,7 +173,7 @@ describe("llm route", () => {
         framing: fakeFraming,
       })
 
-      const prepared = yield* compileRequest(
+      const prepared = yield* (yield* LLMClient.Service).prepare(
         LLMRequest.update(request, { model: updateModel(request.model, { route: duplicate }) }),
       )
 

packages/ai/test/cache-policy.test.ts

@@ -1,8 +1,7 @@
 import { describe, expect, test } from "bun:test"
 import { Effect } from "effect"
 import { CacheHint, LLM, Message } from "../src"
-import { Auth } from "../src/route"
-import { compileRequest } from "../src/route/client"
+import { Auth, LLMClient } from "../src/route"
 import { AmazonBedrock } from "../src/providers"
 import * as AnthropicMessages from "../src/protocols/anthropic-messages"
 import * as Gemini from "../src/protocols/gemini"
@@ -32,7 +31,7 @@ const geminiModel = Gemini.route
 describe("applyCachePolicy", () => {
   it.effect("undefined cache resolves to 'auto' (the recommended default)", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: anthropicModel,
           system: "You are concise.",
@@ -51,7 +50,7 @@ describe("applyCachePolicy", () => {
 
   it.effect("'auto' marks the last tool, first and last system parts, and final message boundary on Anthropic", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: anthropicModel,
           system: [
@@ -88,7 +87,7 @@ describe("applyCachePolicy", () => {
 
   it.effect("'auto' is a no-op on OpenAI (implicit caching protocol)", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: openaiModel,
           system: "Sys",
@@ -107,7 +106,7 @@ describe("applyCachePolicy", () => {
 
   it.effect("'auto' is a no-op on Gemini (out-of-band caching protocol)", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: geminiModel,
           system: "Sys",
@@ -124,7 +123,7 @@ describe("applyCachePolicy", () => {
 
   it.effect("'auto' on Bedrock emits cachePoint markers in the right places", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: bedrockModel,
           system: [
@@ -158,7 +157,7 @@ describe("applyCachePolicy", () => {
 
   it.effect("'none' disables auto placement even when manual hints exist", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: anthropicModel,
           system: "Sys",
@@ -177,7 +176,7 @@ describe("applyCachePolicy", () => {
 
   it.effect("granular object form: tools-only marks just tools", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: anthropicModel,
           system: "Sys",
@@ -196,7 +195,7 @@ describe("applyCachePolicy", () => {
 
   it.effect("auto policy preserves manual CacheHints on other parts", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: anthropicModel,
           system: [
@@ -242,7 +241,7 @@ describe("applyCachePolicy", () => {
       expect("cache" in tail ? tail.cache : undefined).toBeUndefined()
       expect(applyCachePolicy(applied)).toBe(applied)
 
-      const prepared = yield* compileRequest(request)
+      const prepared = yield* LLMClient.prepare(request)
 
       const body = prepared.body as {
         tools: Array<{ cache_control?: unknown }>
@@ -262,7 +261,7 @@ describe("applyCachePolicy", () => {
 
   it.effect("ttlSeconds in the policy flows through to wire markers", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: anthropicModel,
           system: "Sys",
@@ -279,7 +278,7 @@ describe("applyCachePolicy", () => {
 
   it.effect("messages: { tail: 2 } marks the last 2 message boundaries", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: anthropicModel,
           messages: [Message.user("u1"), Message.assistant("a1"), Message.user("u2"), Message.assistant("a2")],
@@ -297,7 +296,7 @@ describe("applyCachePolicy", () => {
 
   it.effect("'latest-assistant' marks the last assistant message", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: anthropicModel,
           messages: [Message.user("u1"), Message.assistant("a1"), Message.user("u2")],

packages/ai/test/exports.test.ts

@@ -19,7 +19,6 @@ import {
   OpenResponses,
 } from "@opencode-ai/ai/protocols"
 import * as AnthropicMessages from "@opencode-ai/ai/protocols/anthropic-messages"
-import { TestLLM } from "@opencode-ai/ai/testing"
 
 describe("public exports", () => {
   test("root exposes app-facing runtime APIs", () => {
@@ -29,7 +28,6 @@ describe("public exports", () => {
     expect(ImageInput.bytes).toBeFunction()
     expect(Provider.make).toBeFunction()
     expect(ProviderSubpath.make).toBe(Provider.make)
-    expect(TestLLM.layer).toBeFunction()
   })
 
   test("route barrel exposes route-authoring APIs", () => {

packages/ai/test/llm-option-types.types.ts

@@ -1,47 +0,0 @@
-import { Schema } from "effect"
-import { LLM, type Model, type ModelProviderOptions, type ProviderOptions } from "../src"
-import { OpenAIChat } from "../src/protocols"
-
-interface ExampleOptions {
-  readonly [key: string]: unknown
-  readonly mode?: "fast" | "thorough"
-}
-
-type ExampleProviderOptions = ProviderOptions & {
-  readonly example?: ExampleOptions
-}
-
-const model = OpenAIChat.route
-  .with({ endpoint: { baseURL: "https://example.com/v1" } })
-  .model<ExampleProviderOptions>({ id: "example" })
-
-LLM.request({ model, prompt: "Hello", providerOptions: { example: { mode: "fast" } } })
-LLM.request({ model, prompt: "Hello", providerOptions: { future: { option: true } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Known provider options preserve their value types.
-  providerOptions: { example: { mode: "slow" } },
-})
-
-LLM.generateObject({
-  model,
-  prompt: "Hello",
-  schema: Schema.Struct({ answer: Schema.String }),
-  providerOptions: { example: { mode: "thorough" } },
-})
-
-LLM.generateObject({
-  model,
-  prompt: "Hello",
-  jsonSchema: { type: "object" },
-  // @ts-expect-error Dynamic object generation uses the selected model's provider options.
-  providerOptions: { example: { mode: false } },
-})
-
-declare const generic: Model
-LLM.request({ model: generic, prompt: "Hello", providerOptions: { arbitrary: { option: true } } })
-
-const options: ModelProviderOptions<typeof model> = { example: { mode: "fast" } }
-void options

packages/ai/test/prepare.test.ts

@@ -4,7 +4,6 @@ import { HttpClientRequest } from "effect/unstable/http"
 import { LLM, mergeProviderOptions } from "../src"
 import { AnthropicMessages, OpenAIChat } from "../src/protocols"
 import { Auth, LLMClient } from "../src/route"
-import { compileRequest } from "../src/route/client"
 import { it } from "./lib/effect"
 import { dynamicResponse } from "./lib/http"
 import { deltaChunk } from "./lib/openai-chunks"
@@ -45,7 +44,7 @@ describe("request option precedence", () => {
     })
   })
 
-  it.effect("compiles bodies with route defaults, model defaults, and call options in order", () =>
+  it.effect("prepares bodies with route defaults, model defaults, and call options in order", () =>
     Effect.gen(function* () {
       const route = OpenAIChat.route.with({
         endpoint: { baseURL: "https://api.openai.test/v1/" },
@@ -60,7 +59,7 @@ describe("request option precedence", () => {
           providerOptions: { openai: { reasoningEffort: "medium" } },
         },
       })
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           prompt: "Say hello.",
@@ -142,7 +141,7 @@ describe("request option precedence", () => {
       const model = OpenAIChat.route
         .with({ endpoint: { baseURL: "https://api.openai.test/v1/" }, auth: Auth.bearer("test") })
         .model({ id: "gpt-4o-mini" })
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           model,
           prompt: "Say hello.",
@@ -165,8 +164,10 @@ describe("request option precedence", () => {
         limits: { output: 128 },
       })
       const model = route.model({ id: "claude-sonnet-4-5", defaults: { limits: { output: 64 } } })
-      const withoutMaxTokens = yield* compileRequest(LLM.request({ model, prompt: "Say hello.", cache: "none" }))
-      const withMaxTokens = yield* compileRequest(
+      const withoutMaxTokens = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
+        LLM.request({ model, prompt: "Say hello.", cache: "none" }),
+      )
+      const withMaxTokens = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLM.request({ model, prompt: "Say hello.", cache: "none", generation: { maxTokens: 32 } }),
       )
 

packages/ai/test/provider-error.test.ts

@@ -58,12 +58,6 @@ describe("provider error classification", () => {
     ).toEqual(["ProviderInternal", "ProviderInternal"])
   })
 
-  test("classifies transient client statuses as provider internal", () => {
-    expect(
-      [408, 409].map((status) => classifyProviderFailure({ message: `HTTP ${status}`, status })._tag),
-    ).toEqual(["ProviderInternal", "ProviderInternal"])
-  })
-
   test("classifies nested provider codes when a top-level code is also present", () => {
     expect(
       [

packages/ai/test/provider-options/anthropic-compatible.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { AnthropicCompatible } from "../../src/providers"
-
-const model = AnthropicCompatible.configure({ baseURL: "https://example.com" }).model("claude")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { anthropic: { effort: "high" } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Anthropic effort must be a string.
-  providerOptions: { anthropic: { effort: 1 } },
-})

packages/ai/test/provider-options/anthropic.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { Anthropic } from "../../src/providers"
-
-const model = Anthropic.provider.model("claude-sonnet-4-5")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { anthropic: { thinking: { type: "adaptive" } } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Anthropic thinking modes are a fixed union.
-  providerOptions: { anthropic: { thinking: { type: "automatic" } } },
-})

packages/ai/test/provider-options/azure.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { Azure } from "../../src/providers"
-
-const model = Azure.configure({ resourceName: "example" }).responses("deployment")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { openai: { store: false } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Azure OpenAI store must be boolean.
-  providerOptions: { openai: { store: "false" } },
-})

packages/ai/test/provider-options/cloudflare.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { CloudflareWorkersAI } from "../../src/providers"
-
-const model = CloudflareWorkersAI.configure({ accountId: "account", apiKey: "test" }).model("model")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { openai: { promptCacheKey: "cache" } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Cloudflare's OpenAI-compatible prompt cache key must be a string.
-  providerOptions: { openai: { promptCacheKey: 1 } },
-})

packages/ai/test/provider-options/github-copilot.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { GitHubCopilot } from "../../src/providers"
-
-const model = GitHubCopilot.configure({ baseURL: "https://example.com" }).model("gpt-5")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { openai: { reasoningSummary: "auto" } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Copilot reasoning summaries use the OpenAI union.
-  providerOptions: { openai: { reasoningSummary: "full" } },
-})

packages/ai/test/provider-options/google-vertex-chat.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { GoogleVertexChat } from "../../src/providers"
-
-const model = GoogleVertexChat.configure({ accessToken: "test", project: "project" }).model("gemini")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { openai: { serviceTier: "priority" } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Vertex OpenAI-compatible service tiers use the OpenAI union.
-  providerOptions: { openai: { serviceTier: "premium" } },
-})

packages/ai/test/provider-options/google-vertex-messages.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { GoogleVertexMessages } from "../../src/providers"
-
-const model = GoogleVertexMessages.configure({ accessToken: "test", project: "project" }).model("claude")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { anthropic: { effort: "medium" } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Vertex Anthropic effort must be a string.
-  providerOptions: { anthropic: { effort: false } },
-})

packages/ai/test/provider-options/google-vertex-responses.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { GoogleVertexResponses } from "../../src/providers"
-
-const model = GoogleVertexResponses.configure({ accessToken: "test", project: "project" }).model("gemini")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { openresponses: { textVerbosity: "high" } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Vertex Responses verbosity uses the Open Responses union.
-  providerOptions: { openresponses: { textVerbosity: "verbose" } },
-})

packages/ai/test/provider-options/google-vertex.types.ts

@@ -1,17 +0,0 @@
-import { LLM } from "../../src"
-import { GoogleVertex } from "../../src/providers"
-
-const model = GoogleVertex.provider.configure({ apiKey: "test" }).model("gemini-2.5-pro")
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  providerOptions: { gemini: { thinkingConfig: { includeThoughts: true } } },
-})
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Vertex Gemini includeThoughts must be boolean.
-  providerOptions: { gemini: { thinkingConfig: { includeThoughts: "yes" } } },
-})

packages/ai/test/provider-options/google.types.ts

@@ -1,17 +0,0 @@
-import { LLM } from "../../src"
-import { Google } from "../../src/providers"
-
-const model = Google.provider.model("gemini-2.5-pro")
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  providerOptions: { gemini: { thinkingConfig: { thinkingBudget: 1024 } } },
-})
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Gemini thinking budgets must be numeric.
-  providerOptions: { gemini: { thinkingConfig: { thinkingBudget: "large" } } },
-})

packages/ai/test/provider-options/openai-compatible-responses.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { OpenAICompatibleResponses } from "../../src/providers"
-
-const model = OpenAICompatibleResponses.configure({ baseURL: "https://example.com" }).model("model")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { openresponses: { reasoningSummary: "detailed" } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error Open Responses reasoning summaries use a fixed union.
-  providerOptions: { openresponses: { reasoningSummary: "full" } },
-})

packages/ai/test/provider-options/openai-compatible.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { OpenAICompatible } from "../../src/providers"
-
-const model = OpenAICompatible.deepseek.model("deepseek-chat")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { openai: { store: false } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error OpenAI-compatible store must be boolean.
-  providerOptions: { openai: { store: "false" } },
-})

packages/ai/test/provider-options/openai.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { OpenAI } from "../../src/providers"
-
-const model = OpenAI.responses("gpt-5")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { openai: { reasoningEffort: "high" } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error OpenAI reasoning effort must be a string.
-  providerOptions: { openai: { reasoningEffort: 1 } },
-})

packages/ai/test/provider-options/openrouter.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { OpenRouter } from "../../src/providers"
-
-const model = OpenRouter.provider.model("anthropic/claude-sonnet-4.5")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { openrouter: { usage: true } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error OpenRouter usage must be boolean or an option record.
-  providerOptions: { openrouter: { usage: "yes" } },
-})

packages/ai/test/provider-options/xai.types.ts

@@ -1,13 +0,0 @@
-import { LLM } from "../../src"
-import { XAI } from "../../src/providers"
-
-const model = XAI.provider.model("grok-4")
-
-LLM.request({ model, prompt: "Hello", providerOptions: { openai: { reasoningEffort: "high" } } })
-
-LLM.request({
-  model,
-  prompt: "Hello",
-  // @ts-expect-error xAI's OpenAI-compatible reasoning effort must be a string.
-  providerOptions: { openai: { reasoningEffort: true } },
-})

packages/ai/test/provider/anthropic-messages.test.ts

@@ -3,7 +3,6 @@ import { Effect } from "effect"
 import { HttpClientRequest } from "effect/unstable/http"
 import { CacheHint, LLM, LLMError, LLMRequest, Message, ToolCallPart, ToolDefinition, Usage } from "../../src"
 import { Auth, LLMClient } from "../../src/route"
-import { compileRequest } from "../../src/route/client"
 import * as AnthropicMessages from "../../src/protocols/anthropic-messages"
 import { continuationRequest, nativeAnthropicMessagesContinuation } from "../continuation-scenarios"
 import { it } from "../lib/effect"
@@ -45,7 +44,7 @@ const expectToolResult = (body: AnthropicMessages.AnthropicMessagesBody): Anthro
 describe("Anthropic Messages route", () => {
   it.effect("prepares Anthropic Messages target", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(request)
+      const prepared = yield* LLMClient.prepare(request)
 
       expect(prepared.body).toEqual({
         model: "claude-sonnet-4-5",
@@ -60,7 +59,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("lowers adaptive thinking settings with effort", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLMRequest.update(request, {
           providerOptions: {
             anthropic: { thinking: { type: "adaptive", display: "summarized" }, effort: "low" },
@@ -77,17 +76,17 @@ describe("Anthropic Messages route", () => {
 
   it.effect("normalizes enabled and disabled thinking settings", () =>
     Effect.gen(function* () {
-      const enabled = yield* compileRequest(
+      const enabled = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLMRequest.update(request, {
           providerOptions: { anthropic: { thinking: { type: "enabled", budgetTokens: 1_024 } } },
         }),
       )
-      const legacy = yield* compileRequest(
+      const legacy = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLMRequest.update(request, {
           providerOptions: { anthropic: { thinking: { type: "enabled", budget_tokens: 2_048 } } },
         }),
       )
-      const disabled = yield* compileRequest(
+      const disabled = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLMRequest.update(request, {
           providerOptions: { anthropic: { thinking: { type: "disabled" } } },
         }),
@@ -101,7 +100,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("rejects enabled thinking without a budget", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLMRequest.update(request, {
           providerOptions: { anthropic: { thinking: { type: "enabled" } } },
         }),
@@ -113,7 +112,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("lowers chronological system updates natively for Claude Opus 4.8 with cache hints", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLM.request({
           model: opus48,
           messages: [
@@ -138,7 +137,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("lowers chronological system updates to wrapped user text for unsupported Anthropic models", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLM.request({
           model,
           messages: [
@@ -165,7 +164,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("rejects non-text chronological system update content before send", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           model: opus48,
           messages: [
@@ -182,7 +181,7 @@ describe("Anthropic Messages route", () => {
   it.effect("falls back for unsupported native chronological system update placement", () =>
     Effect.gen(function* () {
       expect(
-        (yield* compileRequest(
+        (yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
           LLM.request({
             model: opus48,
             messages: [Message.assistant("Plain."), Message.system("After plain assistant.")],
@@ -197,11 +196,12 @@ describe("Anthropic Messages route", () => {
         },
       ])
       expect(
-        (yield* compileRequest(LLM.request({ model: opus48, messages: [Message.system("First.")], cache: "none" })))
-          .body.messages,
+        (yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
+          LLM.request({ model: opus48, messages: [Message.system("First.")], cache: "none" }),
+        )).body.messages,
       ).toEqual([{ role: "user", content: [{ type: "text", text: "<system-update>\nFirst.\n</system-update>" }] }])
       expect(
-        (yield* compileRequest(
+        (yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
           LLM.request({
             model: opus48,
             messages: [Message.user("Before."), Message.system("One."), Message.system("Two.")],
@@ -223,7 +223,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("rejects a system update between a local tool call and its result", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           model: opus48,
           messages: [
@@ -242,7 +242,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("prepares tool call and tool result messages", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLM.request({
           id: "req_tool_result",
           model,
@@ -273,7 +273,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("keeps tools and sends tool_choice none", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLM.request({
           id: "req_tool_choice_none",
           model,
@@ -303,7 +303,7 @@ describe("Anthropic Messages route", () => {
   // not JSON-stringified into `tool_result.content`.
   it.effect("lowers media tool-result content as structured blocks", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLM.request({
           id: "req_tool_result_image",
           model,
@@ -335,7 +335,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("lowers single-image tool-result content as a structured image block", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLM.request({
           id: "req_tool_result_image_only",
           model,
@@ -360,7 +360,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("rejects unsupported media in tool-result content with a clear error", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           id: "req_tool_result_unsupported_media",
           model,
@@ -384,7 +384,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("prepares the composed native continuation request", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         continuationRequest({
           id: "req_native_continuation_anthropic",
           model,
@@ -428,7 +428,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("lowers preserved Anthropic reasoning signature metadata", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model,
           messages: [
@@ -447,7 +447,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("round-trips redacted thinking as redacted_thinking blocks", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model,
           messages: [
@@ -628,7 +628,9 @@ describe("Anthropic Messages route", () => {
         { type: "reasoning", text: "", providerMetadata: { anthropic: { signature: "sig_1" } } },
       ])
 
-      const prepared = yield* compileRequest(LLM.request({ model, messages: [response.message], cache: "none" }))
+      const prepared = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
+        LLM.request({ model, messages: [response.message], cache: "none" }),
+      )
       expect(prepared.body.messages).toEqual([
         { role: "assistant", content: [{ type: "thinking", thinking: "", signature: "sig_1" }] },
       ])
@@ -771,7 +773,7 @@ describe("Anthropic Messages route", () => {
           ),
         ),
       )
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<AnthropicMessages.AnthropicMessagesBody>(
         LLM.request({
           model,
           messages: [
@@ -1131,7 +1133,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("round-trips provider-executed assistant content into server tool blocks", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_round_trip",
           model,
@@ -1182,7 +1184,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("rejects round-trip for unknown server tool names", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           id: "req_unknown_server_tool",
           model,
@@ -1259,7 +1261,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("maps ttlSeconds >= 3600 to cache_control ttl: '1h'", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model,
           system: { type: "text", text: "system", cache: new CacheHint({ type: "ephemeral", ttlSeconds: 3600 }) },
@@ -1275,7 +1277,7 @@ describe("Anthropic Messages route", () => {
 
   it.effect("emits cache_control on tool definitions and tool-result blocks", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model,
           tools: [
@@ -1316,7 +1318,7 @@ describe("Anthropic Messages route", () => {
   it.effect("drops cache_control breakpoints past the 4-per-request cap", () =>
     Effect.gen(function* () {
       const hint = new CacheHint({ type: "ephemeral" })
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model,
           system: [
@@ -1342,7 +1344,7 @@ describe("Anthropic Messages route", () => {
   it.effect("spends breakpoint budget on tools before system before messages", () =>
     Effect.gen(function* () {
       const hint = new CacheHint({ type: "ephemeral" })
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model,
           tools: [

packages/ai/test/provider/bedrock-converse.test.ts

@@ -13,7 +13,6 @@ import {
   ToolDefinition,
 } from "../../src"
 import { LLMClient } from "../../src/route"
-import { compileRequest } from "../../src/route/client"
 import { AmazonBedrock } from "../../src/providers"
 import * as BedrockConverse from "../../src/protocols/bedrock-converse"
 import { it } from "../lib/effect"
@@ -102,7 +101,7 @@ const baseRequest = LLM.request({
 describe("Bedrock Converse route", () => {
   it.effect("prepares Converse target with system, inference config, and messages", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(baseRequest)
+      const prepared = yield* LLMClient.prepare(baseRequest)
 
       expect(prepared.body).toEqual({
         modelId: "anthropic.claude-3-5-sonnet-20240620-v1:0",
@@ -115,7 +114,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("passes topK through additionalModelRequestFields as top_k", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<BedrockConverse.BedrockConverseBody>(
         LLMRequest.update(baseRequest, {
           generation: GenerationOptions.make({ maxTokens: 64, temperature: 0, topK: 40 }),
         }),
@@ -130,14 +129,14 @@ describe("Bedrock Converse route", () => {
 
   it.effect("omits additionalModelRequestFields when topK is unset", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(baseRequest)
+      const prepared = yield* LLMClient.prepare<BedrockConverse.BedrockConverseBody>(baseRequest)
       expect(prepared.body.additionalModelRequestFields).toBeUndefined()
     }),
   )
 
   it.effect("lowers chronological system updates to wrapped user text in order", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<BedrockConverse.BedrockConverseBody>(
         LLM.request({
           model,
           messages: [Message.user("Before."), Message.system("Update."), Message.assistant("After.")],
@@ -154,7 +153,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("prepares tool config with toolSpec and toolChoice", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLMRequest.update(baseRequest, {
           tools: [
             ToolDefinition.make({
@@ -188,7 +187,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("keeps tools and omits the unsupported choice when tool choice is none", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLMRequest.update(baseRequest, {
           tools: [
             ToolDefinition.make({
@@ -218,7 +217,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("lowers assistant tool-call + tool-result message history", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_history",
           model,
@@ -257,7 +256,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("lowers image content in tool-result messages", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_tool_image",
           model,
@@ -492,7 +491,7 @@ describe("Bedrock Converse route", () => {
         providerMetadata: { bedrock: { signature: "sig_1" } },
       })
 
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<BedrockConverse.BedrockConverseBody>(
         LLM.request({
           model,
           messages: [
@@ -547,7 +546,9 @@ describe("Bedrock Converse route", () => {
         },
       ])
 
-      const prepared = yield* compileRequest(LLM.request({ model, messages: [response.message], cache: "none" }))
+      const prepared = yield* LLMClient.prepare<BedrockConverse.BedrockConverseBody>(
+        LLM.request({ model, messages: [response.message], cache: "none" }),
+      )
       expect(prepared.body.messages).toEqual([
         {
           role: "assistant",
@@ -638,7 +639,7 @@ describe("Bedrock Converse route", () => {
         text: "",
         providerMetadata: { bedrock: { redactedData } },
       })
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<BedrockConverse.BedrockConverseBody>(
         LLM.request({
           model,
           messages: [
@@ -753,7 +754,7 @@ describe("Bedrock Converse route", () => {
           secretAccessKey: "wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY",
         },
       }).model("anthropic.claude-3-5-sonnet-20240620-v1:0")
-      const prepared = yield* compileRequest(LLMRequest.update(baseRequest, { model: signed }))
+      const prepared = yield* LLMClient.prepare(LLMRequest.update(baseRequest, { model: signed }))
 
       expect(prepared.route).toBe("bedrock-converse")
       expect(prepared.model).toBe(signed)
@@ -763,7 +764,7 @@ describe("Bedrock Converse route", () => {
   it.effect("emits cachePoint markers after system, user-text, and assistant-text with cache hints", () =>
     Effect.gen(function* () {
       const cache = new CacheHint({ type: "ephemeral" })
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_cache",
           model,
@@ -795,7 +796,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("does not emit cachePoint when no cache hint is set", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(baseRequest)
+      const prepared = yield* LLMClient.prepare(baseRequest)
       expect(prepared.body).toMatchObject({
         system: [{ text: "You are concise." }],
         messages: [{ role: "user", content: [{ text: "Say hello." }] }],
@@ -805,7 +806,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("lowers image media into Bedrock image blocks", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_image",
           model,
@@ -842,7 +843,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("base64-encodes Uint8Array image bytes", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_image_bytes",
           model,
@@ -864,7 +865,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("lowers document media into Bedrock document blocks with format and name", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_doc",
           model,
@@ -896,7 +897,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("requires names for document media", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           model,
           messages: [Message.user({ type: "media", mediaType: "application/pdf", data: "UERGREFUQQ==" })],
@@ -909,7 +910,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("passes named document-only messages through for provider validation", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<BedrockConverse.BedrockConverseBody>(
         LLM.request({
           model,
           cache: "none",
@@ -935,7 +936,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("lowers document media in tool results", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<BedrockConverse.BedrockConverseBody>(
         LLM.request({
           model,
           cache: "none",
@@ -987,7 +988,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("rejects unsupported image media types", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           id: "req_bad_image",
           model,
@@ -1001,7 +1002,7 @@ describe("Bedrock Converse route", () => {
 
   it.effect("rejects unsupported document media types", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           id: "req_bad_doc",
           model,
@@ -1016,7 +1017,7 @@ describe("Bedrock Converse route", () => {
   it.effect("maps ttlSeconds >= 3600 to cachePoint ttl: '1h'", () =>
     Effect.gen(function* () {
       const cache = new CacheHint({ type: "ephemeral", ttlSeconds: 3600 })
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model,
           system: [{ type: "text", text: "system", cache }],
@@ -1033,7 +1034,7 @@ describe("Bedrock Converse route", () => {
   it.effect("appends cachePoint after marked tool definitions and tool-result blocks", () =>
     Effect.gen(function* () {
       const cache = new CacheHint({ type: "ephemeral" })
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model,
           tools: [{ name: "lookup", description: "lookup", inputSchema: { type: "object", properties: {} }, cache }],
@@ -1065,7 +1066,7 @@ describe("Bedrock Converse route", () => {
   it.effect("drops cachePoint markers past the 4-per-request cap", () =>
     Effect.gen(function* () {
       const cache = new CacheHint({ type: "ephemeral" })
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model,
           system: [

packages/ai/test/provider/cloudflare.test.ts

@@ -3,7 +3,7 @@ import { ConfigProvider, Effect, Schema } from "effect"
 import { HttpClientRequest } from "effect/unstable/http"
 import { LLM, LLMEvent } from "../../src"
 import { CloudflareAIGateway, CloudflareWorkersAI } from "../../src/providers/cloudflare"
-import { compileRequest } from "../../src/route/client"
+import { LLMClient } from "../../src/route"
 import { it } from "../lib/effect"
 import { dynamicResponse } from "../lib/http"
 import { sseEvents } from "../lib/sse"
@@ -34,7 +34,7 @@ describe("Cloudflare", () => {
       })
       expect(model.route.endpoint.baseURL).toBe("https://gateway.ai.cloudflare.com/v1/test-account/test-gateway/compat")
 
-      const prepared = yield* compileRequest(LLM.request({ model, prompt: "Say hello." }))
+      const prepared = yield* LLMClient.prepare(LLM.request({ model, prompt: "Say hello." }))
 
       expect(prepared.route).toBe("cloudflare-ai-gateway")
       expect(prepared.body).toMatchObject({
@@ -129,7 +129,7 @@ describe("Cloudflare", () => {
         openai: { reasoningField: "reasoning", reasoningDetails: merged },
       })
 
-      const replay = yield* compileRequest(LLM.request({ model, messages: [response.message] }))
+      const replay = yield* LLMClient.prepare(LLM.request({ model, messages: [response.message] }))
       expect(replay.body.messages).toEqual([
         { role: "assistant", content: "Hello", reasoning: "Thinking", reasoning_details: merged },
       ])
@@ -180,7 +180,7 @@ describe("Cloudflare", () => {
 
   it.effect("allows a fully configured baseURL override", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: CloudflareAIGateway.configure({
             baseURL: "https://gateway.proxy.test/v1/custom/compat",
@@ -208,7 +208,7 @@ describe("Cloudflare", () => {
       })
       expect(model.route.endpoint.baseURL).toBe("https://api.cloudflare.com/client/v4/accounts/test-account/ai/v1")
 
-      const prepared = yield* compileRequest(LLM.request({ model, prompt: "Say hello." }))
+      const prepared = yield* LLMClient.prepare(LLM.request({ model, prompt: "Say hello." }))
 
       expect(prepared.route).toBe("cloudflare-workers-ai")
       expect(prepared.body).toMatchObject({

packages/ai/test/provider/gemini.test.ts

@@ -2,7 +2,6 @@ import { describe, expect } from "bun:test"
 import { Effect } from "effect"
 import { LLM, LLMError, LLMRequest, Message, ToolCallPart, ToolDefinition, Usage } from "../../src"
 import { Auth, LLMClient } from "../../src/route"
-import { compileRequest } from "../../src/route/client"
 import * as Gemini from "../../src/protocols/gemini"
 import { ProviderShared } from "../../src/protocols/shared"
 import { it } from "../lib/effect"
@@ -27,7 +26,7 @@ const request = LLM.request({
 describe("Gemini route", () => {
   it.effect("prepares Gemini target", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(request)
+      const prepared = yield* LLMClient.prepare(request)
 
       expect(prepared.body).toEqual({
         contents: [{ role: "user", parts: [{ text: "Say hello." }] }],
@@ -39,12 +38,12 @@ describe("Gemini route", () => {
 
   it.effect("normalizes Gemini thinking options", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<Gemini.GeminiBody>(
         LLMRequest.update(request, {
           providerOptions: { gemini: { thinkingConfig: { thinkingBudget: 0, includeThoughts: false } } },
         }),
       )
-      const filtered = yield* compileRequest(
+      const filtered = yield* LLMClient.prepare<Gemini.GeminiBody>(
         LLMRequest.update(request, {
           providerOptions: { gemini: { thinkingConfig: { thinkingBudget: "invalid", includeThoughts: false } } },
         }),
@@ -60,7 +59,7 @@ describe("Gemini route", () => {
 
   it.effect("lowers chronological system updates to wrapped user text in order", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<Gemini.GeminiBody>(
         LLM.request({
           model,
           messages: [Message.user("Before."), Message.system("Update."), Message.assistant("After.")],
@@ -76,7 +75,7 @@ describe("Gemini route", () => {
 
   it.effect("prepares multimodal user input and tool history", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_tool_result",
           model,
@@ -144,7 +143,7 @@ describe("Gemini route", () => {
 
   it.effect("continues media tool results as inline model input without base64 text", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<Gemini.GeminiBody>(
         LLM.request({
           model,
           messages: [
@@ -189,7 +188,7 @@ describe("Gemini route", () => {
 
   it.effect("strips matching data URLs to raw base64 inlineData", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<Gemini.GeminiBody>(
         LLM.request({
           model,
           messages: [
@@ -230,7 +229,7 @@ describe("Gemini route", () => {
   ] as const)
     it.effect(`rejects ${name}`, () =>
       Effect.gen(function* () {
-        const error = yield* compileRequest(
+        const error = yield* LLMClient.prepare(
           LLM.request({ model, messages: [Message.user({ type: "media", ...media })] }),
         ).pipe(Effect.flip)
         expect(error.message).toMatch(/does not support|does not match|valid base64/)
@@ -239,7 +238,7 @@ describe("Gemini route", () => {
 
   it.effect("rejects oversized image input", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           model,
           messages: [
@@ -257,7 +256,7 @@ describe("Gemini route", () => {
 
   it.effect("keeps tools and sends function calling mode NONE", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_tool_choice_none",
           model,
@@ -277,7 +276,7 @@ describe("Gemini route", () => {
 
   it.effect("sanitizes integer enums, dangling required, untyped arrays, and scalar object keys", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_schema_patch",
           model,
@@ -458,7 +457,7 @@ describe("Gemini route", () => {
         response.events.findIndex((event) => event.type === "tool-call"),
       )
 
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<Gemini.GeminiBody>(
         LLM.request({
           model,
           messages: [
@@ -692,7 +691,7 @@ describe("Gemini route", () => {
 
   it.effect("rejects unsupported assistant media content", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           id: "req_media",
           model,

packages/ai/test/provider/google-vertex.test.ts

@@ -4,7 +4,6 @@ import { HttpClientRequest } from "effect/unstable/http"
 import { LLM } from "../../src"
 import { GoogleVertex, GoogleVertexChat, GoogleVertexMessages, GoogleVertexResponses } from "../../src/providers"
 import { LLMClient } from "../../src/route"
-import { compileRequest } from "../../src/route/client"
 import { it } from "../lib/effect"
 import { dynamicResponse } from "../lib/http"
 import { deltaChunk, finishChunk } from "../lib/openai-chunks"
@@ -183,7 +182,7 @@ describe("Google Vertex providers", () => {
 
   it.effect("protects the Vertex Messages API version from body overlays", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           model: GoogleVertexMessages.configure({
             accessToken: "vertex-token",

packages/ai/test/provider/openai-chat-reasoning.recorded.test.ts

@@ -5,7 +5,6 @@ import { OpenAIChat } from "../../src/protocols/openai-chat"
 import * as OpenAICompatible from "../../src/providers/openai-compatible"
 import * as OpenRouter from "../../src/providers/openrouter"
 import { LLMClient } from "../../src/route"
-import { compileRequest } from "../../src/route/client"
 import { recordedTests } from "../recorded-test"
 import { expectWeatherToolLoop, goldenWeatherToolLoopRequest, runWeatherToolLoop } from "../recorded-scenarios"
 
@@ -85,7 +84,9 @@ for (const item of cases) {
             ),
           ).toBe(true)
 
-          const replay = yield* compileRequest(LLM.request({ model: item.model, messages: [response.message] }))
+          const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
+            LLM.request({ model: item.model, messages: [response.message] }),
+          )
           expect(replay.body.messages).toMatchObject([
             { role: "assistant", content: response.text, reasoning: response.reasoning },
           ])

packages/ai/test/provider/openai-chat.test.ts

@@ -18,7 +18,6 @@ import * as OpenAI from "../../src/providers/openai"
 import * as OpenAIChat from "../../src/protocols/openai-chat"
 import { ProviderShared } from "../../src/protocols/shared"
 import { Auth, LLMClient } from "../../src/route"
-import { compileRequest } from "../../src/route/client"
 import { it } from "../lib/effect"
 import { dynamicResponse, fixedResponse, truncatedStream } from "../lib/http"
 import { deltaChunk, usageChunk } from "../lib/openai-chunks"
@@ -43,7 +42,11 @@ const request = LLM.request({
 describe("OpenAI Chat route", () => {
   it.effect("prepares OpenAI Chat payload", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(request)
+      // Pass the OpenAIChat payload type so `prepared.body` is statically
+      // typed to the route's native shape — the assertions below read field
+      // names without `unknown` casts.
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(request)
+      const _typed: { readonly model: string; readonly stream: true } = prepared.body
 
       expect(prepared.body).toEqual({
         model: "gpt-4o-mini",
@@ -61,7 +64,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("lowers chronological system updates to escaped user wrappers in order", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           messages: [
@@ -84,7 +87,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("replays canonical reasoning as OpenAI-compatible reasoning_content", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           messages: [
@@ -102,7 +105,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("writes reasoning to a configured custom field on every assistant message", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model: Model.update(model, { compatibility: { reasoningField: "vendor_reasoning" } }),
           messages: [
@@ -128,7 +131,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("rejects reasoning fields that conflict with assistant message fields", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           model: Model.update(model, { compatibility: { reasoningField: "content" } }),
           messages: [Message.assistant([{ type: "reasoning", text: "thinking" }])],
@@ -141,7 +144,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("maps OpenAI provider options to Chat options", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model: OpenAI.configure({ baseURL: "https://api.openai.test/v1/", apiKey: "test" }).chat("gpt-4o-mini"),
           prompt: "think",
@@ -156,7 +159,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("passes through custom OpenAI-compatible reasoning effort strings", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           prompt: "think",
@@ -250,7 +253,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("prepares assistant tool-call and tool-result messages", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_tool_result",
           model,
@@ -288,7 +291,7 @@ describe("OpenAI Chat route", () => {
   it.effect("preserves structured tool errors for the model", () =>
     Effect.gen(function* () {
       const error = { error: { type: "unknown", message: "Tool execution interrupted" } }
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           messages: [
@@ -308,7 +311,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("continues image tool results as vision input without base64 text", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           messages: [
@@ -352,7 +355,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("orders parallel tool responses before one aggregated vision message", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           messages: [
@@ -402,7 +405,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("aggregates consecutive tool images with a following system update", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           messages: [
@@ -443,7 +446,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("appends system updates without replacing multipart user content", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           messages: [
@@ -471,7 +474,7 @@ describe("OpenAI Chat route", () => {
   ] as const)
     it.effect(`rejects ${name}`, () =>
       Effect.gen(function* () {
-        const error = yield* compileRequest(
+        const error = yield* LLMClient.prepare(
           LLM.request({ model, messages: [Message.user({ type: "media", ...media })] }),
         ).pipe(Effect.flip)
         expect(error.message).toMatch(/does not support|does not match|valid base64/)
@@ -480,7 +483,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("rejects oversized image input", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           model,
           messages: [
@@ -498,7 +501,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("prepares raw and data URL image media as vision input", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           id: "req_media",
           model,
@@ -525,7 +528,7 @@ describe("OpenAI Chat route", () => {
 
   it.effect("lowers reasoning-only assistant history", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           id: "req_reasoning",
           model,
@@ -616,7 +619,9 @@ describe("OpenAI Chat route", () => {
           openai: { reasoningField: field },
         })
 
-        const replay = yield* compileRequest(LLM.request({ model, messages: [response.message] }))
+        const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
+          LLM.request({ model, messages: [response.message] }),
+        )
         expect(replay.body.messages).toEqual([{ role: "assistant", content: "Hello", [field]: "thinking" }])
       }
     }),
@@ -642,7 +647,9 @@ describe("OpenAI Chat route", () => {
         openai: { reasoningField: "vendor_reasoning" },
       })
 
-      const replay = yield* compileRequest(LLM.request({ model: custom, messages: [response.message] }))
+      const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
+        LLM.request({ model: custom, messages: [response.message] }),
+      )
       expect(replay.body.messages).toEqual([{ role: "assistant", content: "Hello", vendor_reasoning: "thinking" }])
     }),
   )
@@ -685,7 +692,9 @@ describe("OpenAI Chat route", () => {
         openai: { reasoningField: "reasoning", reasoningDetails: details },
       })
 
-      const replay = yield* compileRequest(LLM.request({ model, messages: [response.message] }))
+      const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
+        LLM.request({ model, messages: [response.message] }),
+      )
       expect(replay.body.messages).toEqual([
         {
           role: "assistant",
@@ -728,7 +737,9 @@ describe("OpenAI Chat route", () => {
         openai: { reasoningDetails: details },
       })
 
-      const replay = yield* compileRequest(LLM.request({ model, messages: [response.message] }))
+      const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
+        LLM.request({ model, messages: [response.message] }),
+      )
       expect(replay.body.messages).toEqual([{ role: "assistant", content: "Hello", reasoning_details: details }])
     }),
   )
@@ -753,7 +764,9 @@ describe("OpenAI Chat route", () => {
         openai: { reasoningField: "reasoning", reasoningDetails: details },
       })
 
-      const replay = yield* compileRequest(LLM.request({ model, messages: [response.message] }))
+      const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
+        LLM.request({ model, messages: [response.message] }),
+      )
       expect(replay.body.messages).toEqual([
         { role: "assistant", content: "Hello", reasoning: "thinking", reasoning_details: details },
       ])
@@ -826,7 +839,9 @@ describe("OpenAI Chat route", () => {
         openai: { reasoningDetails: [] },
       })
 
-      const replay = yield* compileRequest(LLM.request({ model, messages: [response.message] }))
+      const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
+        LLM.request({ model, messages: [response.message] }),
+      )
       expect(replay.body.messages).toEqual([{ role: "assistant", content: "Hello", reasoning_details: [] }])
     }),
   )
@@ -874,7 +889,9 @@ describe("OpenAI Chat route", () => {
         response.events.findIndex(LLMEvent.is.textStart),
       )
 
-      const replay = yield* compileRequest(LLM.request({ model, messages: [response.message] }))
+      const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
+        LLM.request({ model, messages: [response.message] }),
+      )
       expect(replay.body.messages).toEqual([
         { role: "assistant", content: "Hello", reasoning: "thinking", reasoning_details: merged },
       ])
@@ -901,7 +918,9 @@ describe("OpenAI Chat route", () => {
       expect(response.events.filter(LLMEvent.is.reasoningStart)).toHaveLength(1)
       expect(response.events.filter(LLMEvent.is.reasoningEnd)).toHaveLength(1)
 
-      const replay = yield* compileRequest(LLM.request({ model, messages: [response.message] }))
+      const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
+        LLM.request({ model, messages: [response.message] }),
+      )
       expect(replay.body.messages).toEqual([{ role: "assistant", content: null, reasoning_details: details }])
     }),
   )
@@ -931,7 +950,7 @@ describe("OpenAI Chat route", () => {
     Effect.gen(function* () {
       const first = { type: "reasoning.text", text: "first", signature: "signed-0", index: 0 }
       const second = { type: "reasoning.text", text: "second", signature: "signed-1", index: 1 }
-      const replay = yield* compileRequest(
+      const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           messages: [
@@ -960,7 +979,7 @@ describe("OpenAI Chat route", () => {
   it.effect("retains scalar replay for mixed structured reasoning parts", () =>
     Effect.gen(function* () {
       const detail = { type: "reasoning.encrypted", data: "opaque", index: 0 }
-      const replay = yield* compileRequest(
+      const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           messages: [
@@ -985,7 +1004,7 @@ describe("OpenAI Chat route", () => {
   it.effect("replays native scalar reasoning alongside native details", () =>
     Effect.gen(function* () {
       const details = [{ type: "reasoning.encrypted", data: "opaque", index: 0 }]
-      const replay = yield* compileRequest(
+      const replay = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           messages: [

packages/ai/test/provider/openai-compatible-chat.test.ts

@@ -3,7 +3,6 @@ import { Effect, Schema } from "effect"
 import { HttpClientRequest } from "effect/unstable/http"
 import { LLM, LLMRequest, Message, ToolCallPart, ToolChoice, ToolDefinition } from "../../src"
 import { Auth, LLMClient } from "../../src/route"
-import { compileRequest } from "../../src/route/client"
 import * as OpenAICompatible from "../../src/providers/openai-compatible"
 import * as OpenAICompatibleChat from "../../src/protocols/openai-compatible-chat"
 import { it } from "../lib/effect"
@@ -53,7 +52,7 @@ const providerFamilies = [
 describe("OpenAI-compatible Chat route", () => {
   it.effect("prepares generic Chat target", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLMRequest.update(request, {
           tools: [ToolDefinition.make({ name: "lookup", description: "Lookup data", inputSchema: { type: "object" } })],
           toolChoice: ToolChoice.make({ type: "required" }),
@@ -128,7 +127,7 @@ describe("OpenAI-compatible Chat route", () => {
 
   it.effect("matches AI SDK compatible basic request body fixture", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(request)
+      const prepared = yield* LLMClient.prepare(request)
 
       expect(prepared.body).toEqual({
         model: "deepseek-chat",
@@ -146,7 +145,7 @@ describe("OpenAI-compatible Chat route", () => {
 
   it.effect("matches AI SDK compatible tool request body fixture", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_tool_parity",
           model,

packages/ai/test/provider/openai-compatible-responses.test.ts

@@ -7,7 +7,6 @@ import { OpenResponses } from "../../src/protocols/open-responses"
 import { OpenAICompatibleResponses } from "../../src/protocols/openai-compatible-responses"
 import { OpenAIResponses } from "../../src/protocols/openai-responses"
 import { LLMClient } from "../../src/route"
-import { compileRequest } from "../../src/route/client"
 import { it } from "../lib/effect"
 import { fixedResponse } from "../lib/http"
 import { sseEvents } from "../lib/sse"
@@ -24,7 +23,7 @@ describe("Open Responses-compatible route", () => {
         baseURL: "https://responses.example.test/v1",
         provider: "example",
       }).model("example-model")
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model,
           system: "You are concise.",
@@ -62,7 +61,7 @@ describe("Open Responses-compatible route", () => {
         apiKey: "test-key",
         baseURL: "https://responses.example.test/v1",
       }).model("example-model")
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({ model, prompt: "Draw.", tools: [OpenAI.imageGeneration()] }),
       ).pipe(Effect.flip)
 
@@ -77,7 +76,7 @@ describe("Open Responses-compatible route", () => {
         apiKey: "test-key",
         baseURL: "https://responses.example.test/v1",
       }).model("example-model")
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model,
           messages: [
@@ -103,7 +102,7 @@ describe("Open Responses-compatible route", () => {
         baseURL: "https://responses.example.test/v1",
         providerOptions: { openresponses: { reasoningEffort: "low", store: true } },
       }).model("example-model")
-      const prepared = yield* compileRequest(LLM.request({ model, prompt: "Think." }))
+      const prepared = yield* LLMClient.prepare(LLM.request({ model, prompt: "Think." }))
 
       expect(prepared.body).toMatchObject({
         reasoning: { effort: "low" },

packages/ai/test/provider/openai-responses.test.ts

@@ -14,7 +14,6 @@ import {
   Usage,
 } from "../../src"
 import { Auth, LLMClient, RequestExecutor, WebSocketExecutor } from "../../src/route"
-import { compileRequest } from "../../src/route/client"
 import * as Azure from "../../src/providers/azure"
 import * as OpenAI from "../../src/providers/openai"
 import * as XAI from "../../src/providers/xai"
@@ -57,7 +56,7 @@ const expectToolOutput = (body: OpenAIResponses.OpenAIResponsesBody): OpenAITool
 describe("OpenAI Responses route", () => {
   it.effect("prepares OpenAI Responses target", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(request)
+      const prepared = yield* LLMClient.prepare(request)
 
       expect(prepared.body).toEqual({
         model: "gpt-4.1-mini",
@@ -75,7 +74,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("lowers the hosted OpenAI image generation tool", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model,
           prompt: "Show me a rooftop garden.",
@@ -93,7 +92,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("rejects invalid hosted image generation options locally", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           model,
           prompt: "Show me a rooftop garden.",
@@ -110,7 +109,7 @@ describe("OpenAI Responses route", () => {
     Effect.gen(function* () {
       const input = LLMRequest.update(request, { providerOptions: { openai: { serviceTier: "priority" } } })
       expect(input.providerOptions).toEqual({ openai: { serviceTier: "priority" } })
-      const prepared = yield* compileRequest(input)
+      const prepared = yield* LLMClient.prepare(input)
 
       expect(prepared.body).toMatchObject({ service_tier: "priority" })
       expect(prepared.body).not.toHaveProperty("serviceTier")
@@ -119,7 +118,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("passes through custom OpenAI reasoning effort strings", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLMRequest.update(request, { providerOptions: { openai: { reasoningEffort: "experimental" } } }),
       )
 
@@ -129,7 +128,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("omits unsupported semantic service tiers", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLMRequest.update(request, { providerOptions: { openai: { serviceTier: "unsupported" } } }),
       )
 
@@ -139,7 +138,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("flattens top-level object unions in function schemas", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLMRequest.update(request, {
           tools: [
             ToolDefinition.make({
@@ -192,7 +191,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("lowers chronological system updates to escaped user wrappers in order", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model,
           messages: [
@@ -218,7 +217,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("prepares OpenAI Responses WebSocket target", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLMRequest.update(request, {
           model: OpenAIResponses.webSocketRoute
             .with({ endpoint: { baseURL: "https://api.openai.test/v1/" }, auth: Auth.bearer("test") })
@@ -396,7 +395,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("prepares function call and function output input items", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_tool_result",
           model,
@@ -433,7 +432,7 @@ describe("OpenAI Responses route", () => {
         content: [],
         structured: {},
       }
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model,
           messages: [
@@ -454,7 +453,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("keeps primitive tool errors as plain text", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model,
           messages: [
@@ -470,7 +469,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("keeps non-JSON tool errors as plain text", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model,
           messages: [
@@ -488,7 +487,7 @@ describe("OpenAI Responses route", () => {
   // image data is not JSON-stringified into `function_call_output.output`.
   it.effect("lowers image tool-result content as structured input_image items", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           id: "req_tool_result_image",
           model,
@@ -517,7 +516,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("lowers single-image tool-result content as structured input_image array", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           id: "req_tool_result_image_only",
           model,
@@ -541,7 +540,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("lowers PDF tool-result content as structured input_file array", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           id: "req_tool_result_pdf",
           model,
@@ -576,7 +575,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("uses xAI inline file encoding for PDF tool results", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model: xaiModel,
           messages: [
@@ -611,7 +610,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("rejects unsupported media in tool-result content with a clear error", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           id: "req_tool_result_unsupported_media",
           model,
@@ -634,7 +633,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("prepares the composed native continuation request", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         continuationRequest({
           id: "req_native_continuation_openai",
           model,
@@ -676,7 +675,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("maps OpenAI provider options to Responses options", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model: OpenAI.configure({ baseURL: "https://api.openai.test/v1/", apiKey: "test" }).model("gpt-5.2"),
           prompt: "think",
@@ -701,7 +700,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("accepts the full ResponseIncludable union", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model,
           prompt: "hi",
@@ -723,7 +722,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("filters unknown includable values out of the include array", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model,
           prompt: "hi",
@@ -740,7 +739,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("treats an explicit empty include as no include at all", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({ model, prompt: "hi", providerOptions: { openai: { include: [] } } }),
       )
 
@@ -750,7 +749,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("treats an all-invalid include as no include at all", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({ model, prompt: "hi", providerOptions: { openai: { include: ["bogus.thing"] } } }),
       )
 
@@ -760,7 +759,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("omits include when no include is set", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({ model, prompt: "hi", providerOptions: { openai: { store: false } } }),
       )
 
@@ -774,7 +773,7 @@ describe("OpenAI Responses route", () => {
       // reasoningSummary: "auto" by default. Without `include`, a follow-up
       // turn cannot replay reasoning state, so the facade also opts into
       // `reasoning.encrypted_content` automatically.
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model: OpenAI.configure({ baseURL: "https://api.openai.test/v1/", apiKey: "test" }).responses("gpt-5.2"),
           prompt: "hi",
@@ -789,7 +788,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("lets callers opt out of the GPT-5 default include", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model: OpenAI.configure({ baseURL: "https://api.openai.test/v1/", apiKey: "test" }).responses("gpt-5.2"),
           prompt: "hi",
@@ -803,7 +802,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("request OpenAI provider options override route defaults", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model: OpenAI.configure({
             baseURL: "https://api.openai.test/v1/",
@@ -935,7 +934,9 @@ describe("OpenAI Responses route", () => {
         },
       ])
 
-      const prepared = yield* compileRequest(LLM.request({ model, messages: [response.message] }))
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
+        LLM.request({ model, messages: [response.message] }),
+      )
       expect(prepared.body.input).toEqual([
         {
           role: "assistant",
@@ -1269,7 +1270,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("preserves assistant content order around reasoning items", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           id: "req_reasoning_order",
           model,
@@ -1307,7 +1308,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("references stored reasoning items by id", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model,
           messages: [
@@ -1329,7 +1330,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("references stored provider-executed hosted tool results by id", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model,
           messages: [
@@ -1366,7 +1367,7 @@ describe("OpenAI Responses route", () => {
   it.effect("continues stateless hosted image generation with the generated image", () =>
     Effect.gen(function* () {
       const imageTool = OpenAI.imageGeneration({ action: "edit" })
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model,
           messages: [
@@ -1407,7 +1408,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("joins streamed summary blocks into one continuation reasoning item", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           id: "req_multi_summary_continuation",
           model,
@@ -1444,7 +1445,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("skips non-persisted reasoning ids without encrypted state", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           id: "req_reasoning_without_encrypted_state",
           model,
@@ -1761,7 +1762,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("lowers user image and PDF content", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           id: "req_media",
           model,
@@ -1792,7 +1793,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("uses xAI inline file encoding for user PDFs", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIResponses.OpenAIResponsesBody>(
         LLM.request({
           model: xaiModel,
           messages: [
@@ -1824,7 +1825,7 @@ describe("OpenAI Responses route", () => {
 
   it.effect("rejects unsupported user media content", () =>
     Effect.gen(function* () {
-      const error = yield* compileRequest(
+      const error = yield* LLMClient.prepare(
         LLM.request({
           id: "req_media",
           model,

packages/ai/test/provider/openrouter.test.ts

@@ -2,7 +2,6 @@ import { describe, expect } from "bun:test"
 import { Effect } from "effect"
 import { LLM, Message } from "../../src"
 import { LLMClient } from "../../src/route"
-import { compileRequest } from "../../src/route/client"
 import * as OpenRouter from "../../src/providers/openrouter"
 import { it } from "../lib/effect"
 import { fixedResponse } from "../lib/http"
@@ -20,7 +19,7 @@ describe("OpenRouter", () => {
       })
       expect(model.route.endpoint.baseURL).toBe("https://openrouter.ai/api/v1")
 
-      const prepared = yield* compileRequest(LLM.request({ model, prompt: "Say hello." }))
+      const prepared = yield* LLMClient.prepare(LLM.request({ model, prompt: "Say hello." }))
 
       expect(prepared.route).toBe("openrouter")
       expect(prepared.body).toMatchObject({
@@ -33,7 +32,7 @@ describe("OpenRouter", () => {
 
   it.effect("applies OpenRouter payload options from the model helper", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare(
         LLM.request({
           model: OpenRouter.configure({
             apiKey: "test-key",
@@ -101,7 +100,7 @@ describe("OpenRouter", () => {
         { type: "reasoning.text", signature: "signed", format: "anthropic-claude-v1", index: 0 },
         { type: "reasoning.encrypted", data: "opaque", format: "openai-responses-v1", index: 1 },
       ]
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenRouter.OpenRouterBody>(
         LLM.request({
           model: OpenRouter.configure({ apiKey: "test-key" }).model("anthropic/claude-sonnet-4.6"),
           messages: [
@@ -134,7 +133,7 @@ describe("OpenRouter", () => {
         { type: "reasoning.encrypted", id: "state", data: "opaque" },
         { type: "reasoning.encrypted", id: "state", data: "opaque" },
       ]
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenRouter.OpenRouterBody>(
         LLM.request({
           model: OpenRouter.configure({ apiKey: "test-key" }).model("anthropic/claude-sonnet-4.6"),
           messages: [
@@ -159,7 +158,7 @@ describe("OpenRouter", () => {
         { type: "reasoning.text", id: "first", index: 0, text: "A", opaque: "first" },
         { type: "reasoning.text", id: "second", index: 1, text: "B", opaque: "second" },
       ]
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenRouter.OpenRouterBody>(
         LLM.request({
           model: OpenRouter.configure({ apiKey: "test-key" }).model("anthropic/claude-sonnet-4.6"),
           messages: [
@@ -180,7 +179,7 @@ describe("OpenRouter", () => {
 
   it.effect("omits scalar reasoning without continuation details", () =>
     Effect.gen(function* () {
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenRouter.OpenRouterBody>(
         LLM.request({
           model: OpenRouter.configure({ apiKey: "test-key" }).model("anthropic/claude-sonnet-4.6"),
           messages: [Message.assistant({ type: "reasoning", text: "Thinking" })],

packages/ai/test/tool-schema-projection.test.ts

@@ -3,8 +3,7 @@ import { Effect } from "effect"
 import { LLM } from "../src"
 import { OpenAIChat } from "../src/protocols"
 import { ToolSchemaProjection } from "../src/protocols/utils/tool-schema"
-import { Auth } from "../src/route"
-import { compileRequest } from "../src/route/client"
+import { Auth, LLMClient } from "../src/route"
 import { it } from "./lib/effect"
 
 describe("tool schema projections", () => {
@@ -80,7 +79,7 @@ describe("tool schema projections", () => {
       const model = OpenAIChat.route
         .with({ endpoint: { baseURL: "https://api.openai.test/v1/" }, auth: Auth.bearer("test") })
         .model({ id: "kimi-k2", compatibility: { toolSchema: "moonshot" } })
-      const prepared = yield* compileRequest(
+      const prepared = yield* LLMClient.prepare<OpenAIChat.OpenAIChatBody>(
         LLM.request({
           model,
           prompt: "Use the tool.",

packages/app/src/components/dialog-fork.tsx

@@ -69,7 +69,7 @@ export const DialogFork: Component = () => {
     const dir = base64Encode(sdk().directory)
 
     sdk()
-      .api.session.fork({ sessionID, boundary: { type: "before", messageID: item.id } })
+      .api.session.fork({ sessionID, messageID: item.id })
       .then((forked) => {
         dialog.close()
         prompt.set(restored, undefined, { dir, id: forked.id })

packages/app/src/components/dialog-select-server.tsx

@@ -562,7 +562,6 @@ export function useServerManagementController(options: { onSelect?: () => void;
     startEdit,
     resetForm,
     submitForm,
-    canRemove: server.canRemove,
     handleRemove,
     handleFormChange: () => (isAddMode() ? handleAddChange : handleEditChange),
     handleFormNameChange: () => (isAddMode() ? handleAddNameChange : handleEditNameChange),
@@ -650,15 +649,13 @@ export function ServerConnectionList(props: { controller: ReturnType<typeof useS
                             </DropdownMenu.ItemLabel>
                           </DropdownMenu.Item>
                         </Show>
-                        <Show when={props.controller.canRemove(key)}>
-                          <DropdownMenu.Separator />
-                          <DropdownMenu.Item
-                            onSelect={() => props.controller.handleRemove(ServerConnection.key(i))}
-                            class="text-text-on-critical-base hover:bg-surface-critical-weak"
-                          >
-                            <DropdownMenu.ItemLabel>{language.t("dialog.server.menu.delete")}</DropdownMenu.ItemLabel>
-                          </DropdownMenu.Item>
-                        </Show>
+                        <DropdownMenu.Separator />
+                        <DropdownMenu.Item
+                          onSelect={() => props.controller.handleRemove(ServerConnection.key(i))}
+                          class="text-text-on-critical-base hover:bg-surface-critical-weak"
+                        >
+                          <DropdownMenu.ItemLabel>{language.t("dialog.server.menu.delete")}</DropdownMenu.ItemLabel>
+                        </DropdownMenu.Item>
                       </DropdownMenu.Content>
                     </DropdownMenu.Portal>
                   </DropdownMenu>

packages/app/src/components/server/server-row-menu.tsx

@@ -21,7 +21,6 @@ export const ServerRowMenu: Component<{
       labels={serverMenuLabels(language)}
       canDefault={props.controller.canDefault()}
       isDefault={props.controller.defaultKey() === key}
-      canRemove={props.controller.canRemove(key)}
       onEdit={props.onEdit}
       onSetDefault={() => props.controller.setDefault(key)}
       onRemoveDefault={() => props.controller.setDefault(null)}
@@ -48,7 +47,6 @@ export const ServerRowMenuView: Component<{
   labels: ReturnType<typeof serverMenuLabels>
   canDefault: boolean
   isDefault: boolean
-  canRemove: boolean
   onEdit: (server: ServerConnection.Http) => void
   onSetDefault: () => void
   onRemoveDefault: () => void
@@ -86,10 +84,10 @@ export const ServerRowMenuView: Component<{
             <Show when={props.canDefault && props.isDefault}>
               <MenuV2.Item onSelect={props.onRemoveDefault}>{props.labels.defaultRemove}</MenuV2.Item>
             </Show>
-            <Show when={props.canRemove}>
-              <MenuV2.Separator />
-              <MenuV2.Item onSelect={props.onRemove}>{props.labels.delete}</MenuV2.Item>
-            </Show>
+            <MenuV2.Separator />
+            <MenuV2.Item disabled={builtin()} onSelect={props.onRemove}>
+              {props.labels.delete}
+            </MenuV2.Item>
           </MenuV2.Group>
         </MenuV2.Content>
       </MenuV2.Portal>

packages/app/src/components/titlebar.tsx

@@ -395,25 +395,27 @@ export function Titlebar(props: { update?: TitlebarUpdate; debugTools?: { visibl
                   }}
                   onReorder={(keys) => tabsStoreActions.reorder(keys)}
                 />
-                <TooltipV2
-                  placement="bottom"
-                  value={
-                    <>
-                      {language.t("command.session.new")}
-                      <KeybindV2 keys={newTabTooltipKeybind(command)} variant="neutral" />
-                    </>
-                  }
-                >
-                  <IconButtonV2
-                    type="button"
-                    variant="ghost-muted"
-                    size="large"
-                    class="shrink-0"
-                    icon={<IconV2 name="plus" />}
-                    onClick={openNewTab}
-                    aria-label={language.t("command.session.new")}
-                  />
-                </TooltipV2>
+                <Show when={!creating()}>
+                  <TooltipV2
+                    placement="bottom"
+                    value={
+                      <>
+                        {language.t("command.session.new")}
+                        <KeybindV2 keys={newTabTooltipKeybind(command)} variant="neutral" />
+                      </>
+                    }
+                  >
+                    <IconButtonV2
+                      type="button"
+                      variant="ghost-muted"
+                      size="large"
+                      class="shrink-0"
+                      icon={<IconV2 name="plus" />}
+                      onClick={openNewTab}
+                      aria-label={language.t("command.session.new")}
+                    />
+                  </TooltipV2>
+                </Show>
                 <div class="flex-1" />
                 <TitlebarV2Right state={v2RightState()} />
               </div>

packages/app/src/context/server.tsx

@@ -178,17 +178,6 @@ export function resolveServerList(input: {
   return [...deduped.values()]
 }
 
-export function canRemoveServer(input: {
-  key: ServerConnection.Key
-  provided?: Array<ServerConnection.Any>
-  stored: StoredServer[]
-}) {
-  if (input.provided?.some((server) => ServerConnection.key(server) === input.key)) return false
-  return input.stored.some((server) =>
-    typeof server === "string" ? server === input.key : ("type" in server ? server.http.url : server.url) === input.key,
-  )
-}
-
 export namespace ServerConnection {
   type Base = { displayName?: string; label?: string }
 
@@ -323,10 +312,6 @@ export const { use: useServer, provider: ServerProvider } = createSimpleContext(
       })
     }
 
-    function canRemove(key: ServerConnection.Key) {
-      return canRemoveServer({ key, provided: props.servers, stored: store.list })
-    }
-
     const isReady = Object.assign(
       createMemo(() => ready() && !!state.active),
       { promise: ready.promise },
@@ -365,7 +350,6 @@ export const { use: useServer, provider: ServerProvider } = createSimpleContext(
       setActive,
       add,
       remove,
-      canRemove,
       scope,
       projects: {
         ...projects,

packages/app/src/pages/home/home-projects-controller.tsx

@@ -60,7 +60,6 @@ export function createHomeProjectsController(home: HomeController) {
       defaultKey: serverManagement.defaultKey,
       setDefault: (conn: ServerConnection.Any | undefined) =>
         serverManagement.setDefault(conn ? ServerConnection.key(conn) : null),
-      canRemove: (conn: ServerConnection.Any) => serverManagement.canRemove(ServerConnection.key(conn)),
       remove: (conn: ServerConnection.Any) => serverManagement.handleRemove(ServerConnection.key(conn)),
       edit: (conn: ServerConnection.Http) => dialog.show(() => <DialogServerV2 mode="edit" server={conn} />),
       focus: home.selection.focusServer,

packages/app/src/pages/home/home-projects-view.tsx

@@ -47,7 +47,6 @@ export type HomeProjectsViewProps = {
   onToggleCollapsed: (server: ServerConnection.Any) => void
   onEditServer: (server: ServerConnection.Http) => void
   onSetDefaultServer: (server: ServerConnection.Any | undefined) => void
-  canRemoveServer: (server: ServerConnection.Any) => boolean
   onRemoveServer: (server: ServerConnection.Any) => void
   onMoveProject: (server: ServerConnection.Any, worktree: string, index: number) => void
   onSelectProject: (server: ServerConnection.Any, directory: string) => void
@@ -193,7 +192,6 @@ function HomeServerRow(props: {
   onToggleCollapsed: HomeProjectsViewProps["onToggleCollapsed"]
   onEditServer: HomeProjectsViewProps["onEditServer"]
   onSetDefaultServer: HomeProjectsViewProps["onSetDefaultServer"]
-  canRemoveServer: HomeProjectsViewProps["canRemoveServer"]
   onRemoveServer: HomeProjectsViewProps["onRemoveServer"]
   onSetContextMenuOpen: HomeProjectsContextMenuProps["onSetContextMenuOpen"]
   onChooseProject: HomeProjectsViewProps["onChooseProject"]
@@ -279,7 +277,6 @@ function HomeServerRow(props: {
           labels={serverMenuLabels(props.language)}
           canDefault={props.canDefaultServer()}
           isDefault={props.defaultServerKey() === ServerConnection.key(props.server)}
-          canRemove={props.canRemoveServer(props.server)}
           onEdit={props.onEditServer}
           onSetDefault={() => props.onSetDefaultServer(props.server)}
           onRemoveDefault={() => props.onSetDefaultServer(undefined)}

packages/app/src/pages/home/home-projects.tsx

@@ -24,7 +24,6 @@ export function HomeProjects(props: { projects: HomeProjectsController; scroll:
       onToggleCollapsed={props.projects.server.toggleCollapsed}
       onEditServer={props.projects.server.edit}
       onSetDefaultServer={props.projects.server.setDefault}
-      canRemoveServer={props.projects.server.canRemove}
       onRemoveServer={props.projects.server.remove}
       onMoveProject={props.projects.project.move}
       onSelectProject={props.projects.project.select}

packages/app/src/utils/server-compat.test.ts

@@ -26,6 +26,7 @@ function setup(
         return new Response(undefined, { status: 204 })
       if (request.method === "POST" && request.url.endsWith("/prompt")) {
         return Response.json({
+          admittedSeq: 1,
           id: "msg_1",
           sessionID: "ses_1",
           timeCreated: 1,
@@ -52,7 +53,6 @@ function setup(
 }
 
 describe("createCompatibleApi", () => {
-  /*
   test("routes V1 archive through the legacy session update", async () => {
     const { api, requests } = setup("v1")
     await api.session.archive({ sessionID: "ses_1", directory: "/repo" })
@@ -63,7 +63,6 @@ describe("createCompatibleApi", () => {
     expect(requests[0]!.method).toBe("PATCH")
     expect(await requests[0]!.json()).toMatchObject({ time: { archived: expect.any(Number) } })
   })
-  */
 
   test("converts current prompts to the V1 prompt contract", async () => {
     const { api, requests } = setup("v1")
@@ -146,7 +145,6 @@ describe("createCompatibleApi", () => {
     expect(detections).toBe(1)
   })
 
-  /*
   test("keeps V2 session actions on the current API", async () => {
     const { api, requests } = setup("v2")
     await api.session.archive({ sessionID: "ses_1" })
@@ -154,7 +152,6 @@ describe("createCompatibleApi", () => {
     expect(new URL(requests[0]!.url).pathname).toBe("/api/session/ses_1/archive")
     expect(requests[0]!.method).toBe("POST")
   })
-  */
 
   test("uses the global V1 session search endpoint", async () => {
     const { api, requests } = setup("v1")

packages/app/src/utils/server-compat.ts

@@ -27,7 +27,7 @@ type CompatibleSessionApi = Omit<
   shell: (input: SessionShellInput & LegacyPrompt) => Promise<SessionShellOutput>
   compact: (input: SessionCompactInput & { model?: LegacyPrompt["model"] }) => Promise<SessionCompactOutput>
   rename: (input: Parameters<SessionApi["rename"]>[0] & LegacyLocation) => ReturnType<SessionApi["rename"]>
-  // archive: (input: Parameters<SessionApi["archive"]>[0] & LegacyLocation) => ReturnType<SessionApi["archive"]>
+  archive: (input: Parameters<SessionApi["archive"]>[0] & LegacyLocation) => ReturnType<SessionApi["archive"]>
   remove: (input: Parameters<SessionApi["remove"]>[0] & LegacyLocation) => ReturnType<SessionApi["remove"]>
 }
 type CompatiblePermissionApi = Omit<ServerApi["permission"], "reply"> & {
@@ -183,9 +183,9 @@ function createV1Api(input: CompatibleInput): CompatibleApi {
       async rename(value: Parameters<ServerApi["session"]["rename"]>[0] & LegacyLocation) {
         await legacy(value).session.update({ sessionID: value.sessionID, title: value.title })
       },
-      // async archive(value: Parameters<ServerApi["session"]["archive"]>[0] & LegacyLocation) {
-      //   await legacy(value).session.update({ sessionID: value.sessionID, time: { archived: Date.now() } })
-      // },
+      async archive(value: Parameters<ServerApi["session"]["archive"]>[0] & LegacyLocation) {
+        await legacy(value).session.update({ sessionID: value.sessionID, time: { archived: Date.now() } })
+      },
       async remove(value: Parameters<ServerApi["session"]["remove"]>[0] & LegacyLocation) {
         await legacy(value).session.delete(value)
       },
@@ -229,6 +229,7 @@ function createV1Api(input: CompatibleInput): CompatibleApi {
           ],
         })
         return {
+          admittedSeq: 0,
           id: value.id ?? "",
           sessionID: value.sessionID,
           timeCreated: Date.now(),
@@ -254,6 +255,7 @@ function createV1Api(input: CompatibleInput): CompatibleApi {
           })),
         })
         return {
+          admittedSeq: 0,
           id: value.id ?? "",
           sessionID: value.sessionID,
           timeCreated: Date.now(),
@@ -278,6 +280,7 @@ function createV1Api(input: CompatibleInput): CompatibleApi {
           modelID: value.model.modelID,
         })
         return {
+          admittedSeq: 0,
           id: value.id ?? "",
           sessionID: value.sessionID,
           timeCreated: Date.now(),

packages/cli/package.json

@@ -56,16 +56,6 @@
     "@lydell/node-pty-linux-x64": "1.2.0-beta.12",
     "@lydell/node-pty-win32-arm64": "1.2.0-beta.12",
     "@lydell/node-pty-win32-x64": "1.2.0-beta.12",
-    "@ff-labs/fff-bin-darwin-arm64": "0.10.1",
-    "@ff-labs/fff-bin-linux-arm64-gnu": "0.10.1",
-    "@ff-labs/fff-bin-linux-x64-gnu": "0.10.1",
-    "@ff-labs/fff-bin-win32-arm64": "0.10.1",
-    "@ff-labs/fff-bin-win32-x64": "0.10.1",
-    "@yuuang/ffi-rs-darwin-arm64": "1.3.2",
-    "@yuuang/ffi-rs-linux-arm64-gnu": "1.3.2",
-    "@yuuang/ffi-rs-linux-x64-gnu": "1.3.2",
-    "@yuuang/ffi-rs-win32-arm64-msvc": "1.3.2",
-    "@yuuang/ffi-rs-win32-x64-msvc": "1.3.2",
     "@parcel/watcher-darwin-arm64": "2.5.1",
     "@parcel/watcher-linux-arm64-glibc": "2.5.1",
     "@parcel/watcher-linux-x64-glibc": "2.5.1",

packages/cli/script/build.ts

@@ -5,7 +5,6 @@ import { rm } from "fs/promises"
 import path from "path"
 import { Script } from "@opencode-ai/script"
 import { createSolidTransformPlugin } from "@opentui/solid/bun-plugin"
-import type { BunPlugin } from "bun"
 import pkg from "../package.json"
 import { modelsData } from "./generate"
 
@@ -23,7 +22,7 @@ await rm(outdir, { recursive: true, force: true })
 const singleFlag = process.argv.includes("--single")
 const baselineFlag = process.argv.includes("--baseline")
 const skipInstall = process.argv.includes("--skip-install")
-const solidPlugin = createSolidTransformPlugin()
+const plugin = createSolidTransformPlugin()
 
 const allTargets: {
   os: string
@@ -56,16 +55,6 @@ const targets = singleFlag
 if (!skipInstall) await $`bun install --os="*" --cpu="*" @opentui/core@${pkg.dependencies["@opentui/core"]}`
 
 for (const item of targets) {
-  const parcelWatcherPackage = `@parcel/watcher-${item.os}-${item.arch}${item.os === "linux" ? `-${item.abi ?? "glibc"}` : ""}`
-  const parcelWatcherPlugin: BunPlugin = {
-    name: "parcel-watcher-binding",
-    setup(build) {
-      build.onLoad({ filter: /filesystem\/watcher-binding\.ts$/ }, () => ({
-        contents: `import binding from ${JSON.stringify(parcelWatcherPackage)}; export default () => binding`,
-        loader: "js",
-      }))
-    },
-  }
   const target = [
     binary,
     item.os === "win32" ? "windows" : item.os,
@@ -80,7 +69,7 @@ for (const item of targets) {
   const result = await Bun.build({
     entrypoints: ["./src/index.ts"],
     tsconfig: "./tsconfig.json",
-    plugins: [solidPlugin, parcelWatcherPlugin],
+    plugins: [plugin],
     external: ["node-gyp"],
     format: "esm",
     minify: true,

packages/cli/script/node-assets.ts

@@ -37,8 +37,6 @@ export async function collectNodeAssets(target: NodeTarget) {
       ...(target.platform === "linux" ? { libc: "glibc" as const } : {}),
     }),
     { key: target.parcelWatcherAsset, source: fileURLToPath(import.meta.resolve(target.parcelWatcherPackage)) },
-    { key: target.fffAsset, source: fileURLToPath(import.meta.resolve(target.fffPackage)) },
-    { key: target.fffFfiAsset, source: fileURLToPath(import.meta.resolve(target.fffFfiPackage)) },
     {
       key: photonWasmAsset,
       source: fileURLToPath(import.meta.resolve(photonWasmAsset)),

packages/cli/script/service-smoke.ts

@@ -13,7 +13,7 @@ const directory = path.join(import.meta.dir, "..", "dist", ...(nodeBuild ? ["nod
 const binary = path.join(directory, `opencode2${nodeBuild ? "-node" : ""}${process.platform === "win32" ? ".exe" : ""}`)
 if (!(await Bun.file(binary).exists())) throw new Error(`Missing compiled CLI in ${directory}`)
 
-const root = await fs.realpath(await fs.mkdtemp(path.join(os.tmpdir(), "opencode-service-smoke-")))
+const root = await fs.mkdtemp(path.join(os.tmpdir(), "opencode-service-smoke-"))
 const env = {
   ...process.env,
   HOME: root,
@@ -29,7 +29,6 @@ const processes: Array<ReturnType<typeof Bun.spawn>> = []
 const errors: Array<Promise<string>> = []
 let failure: unknown
 try {
-  await fs.mkdir(path.join(root, ".opencode"))
   spawnService()
   spawnService()
   const registration = await waitForRegistration()
@@ -50,11 +49,6 @@ try {
     { signal: AbortSignal.timeout(5_000) },
   )
   if (tokenOpenApi.status !== 200) throw new Error("Compiled application rejected query authentication")
-  if ((await pluginIDs(info.url, headers)).includes("smoke")) throw new Error("Smoke plugin existed before creation")
-  const plugin = path.join(root, ".opencode", "plugins", "smoke.ts")
-  await fs.mkdir(path.dirname(plugin), { recursive: true })
-  await fs.writeFile(plugin, pluginSource())
-  await waitForPlugin(info.url, headers)
 
   const unauthorizedHealth = await fetch(new URL("/api/health", info.url), {
     signal: AbortSignal.timeout(5_000),
@@ -94,7 +88,6 @@ try {
 } finally {
   processes.forEach((process) => process.kill())
   await Promise.all(processes.map((process) => process.exited))
-  if (failure) errors.push(fs.readFile(path.join(root, "data", "opencode", "log", "opencode.log"), "utf8").catch(() => ""))
 }
 
 const output = await Promise.all(errors)
@@ -140,31 +133,3 @@ async function waitForReady(url: string, headers: HeadersInit) {
 function exitsWithin(process: Bun.Subprocess, milliseconds: number) {
   return Promise.race([process.exited.then(() => true), Bun.sleep(milliseconds).then(() => false)])
 }
-
-function pluginSource() {
-  return 'export default { id: "smoke", setup: async () => {} }\n'
-}
-
-async function pluginIDs(url: string, headers: HeadersInit) {
-  const endpoint = new URL("/api/plugin", url)
-  endpoint.searchParams.set("location[directory]", root)
-  const response = await fetch(endpoint, { headers, signal: AbortSignal.timeout(5_000) })
-  const body: unknown = await response.json()
-  if (typeof body !== "object" || body === null || !("data" in body) || !Array.isArray(body.data)) {
-    throw new Error("Compiled service returned an invalid plugin list")
-  }
-  return body.data.flatMap((plugin) =>
-    typeof plugin === "object" && plugin !== null && "id" in plugin && typeof plugin.id === "string"
-      ? [plugin.id]
-      : [],
-  )
-}
-
-async function waitForPlugin(url: string, headers: HeadersInit) {
-  const deadline = Date.now() + 10_000
-  while (Date.now() < deadline) {
-    if ((await pluginIDs(url, headers)).includes("smoke")) return
-    await Bun.sleep(25)
-  }
-  throw new Error("Compiled service did not discover the created plugin")
-}

packages/cli/src/acp/service.ts

@@ -244,10 +244,7 @@ export function make(input: { readonly client: OpenCodeClient; readonly connecti
       return {}
     },
     forkSession: async (params) => {
-      const forked = await input.client.session.fork({
-        sessionID: params.sessionId,
-        boundary: { type: "through" },
-      })
+      const forked = await input.client.session.fork({ sessionID: params.sessionId })
       const state = await attach(forked, forked.location.directory, params.mcpServers ?? [])
       await replay(state)
       return { sessionId: state.id, configOptions: configOptions(state) }

packages/cli/src/commands/commands.ts

@@ -14,23 +14,10 @@ const ServerParams = {
   ),
 }
 
-const PermissionParams = {
-  auto: Flag.boolean("auto").pipe(
-    Flag.withDescription("Auto-approve permissions that are not explicitly denied"),
-    Flag.withDefault(false),
-  ),
-  yolo: Flag.boolean("yolo").pipe(Flag.withDefault(false), Flag.withHidden),
-  dangerouslySkipPermissions: Flag.boolean("dangerously-skip-permissions").pipe(
-    Flag.withDefault(false),
-    Flag.withHidden,
-  ),
-}
-
 export const Commands = Spec.make(typeof OPENCODE_CLI_NAME === "string" ? OPENCODE_CLI_NAME : "opencode", {
   description: "OpenCode 2.0 preview command line interface",
   params: {
     ...ServerParams,
-    ...PermissionParams,
     directory: Argument.string("directory").pipe(
       Argument.withDescription("Directory to start OpenCode in"),
       Argument.optional,
@@ -209,7 +196,11 @@ export const Commands = Spec.make(typeof OPENCODE_CLI_NAME === "string" ? OPENCO
         ),
         title: Flag.string("title").pipe(Flag.withDescription("Session title"), Flag.optional),
         thinking: Flag.boolean("thinking").pipe(Flag.withDescription("Show thinking blocks"), Flag.withDefault(false)),
-        ...PermissionParams,
+        auto: Flag.boolean("auto").pipe(
+          Flag.withDescription("Auto-approve permissions that are not explicitly denied"),
+          Flag.withDefault(false),
+        ),
+        yolo: Flag.boolean("yolo").pipe(Flag.withDefault(false), Flag.withHidden),
       },
     }),
     Spec.make("service", {

packages/cli/src/commands/handlers/default.ts

@@ -59,7 +59,6 @@ export default Runtime.handler(Commands, (input) =>
         continue: input.continue,
         sessionID: Option.getOrUndefined(input.session),
         prompt: Option.getOrUndefined(input.prompt),
-        auto: input.auto || input.yolo || input.dangerouslySkipPermissions,
       },
       config: {
         path: config.path,

packages/cli/src/commands/handlers/run.ts

@@ -24,7 +24,7 @@ export default Runtime.handler(Commands.commands.run, (input) =>
         file: [...input.file],
         title: Option.getOrUndefined(input.title),
         thinking: input.thinking,
-        auto: input.auto || input.yolo || input.dangerouslySkipPermissions,
+        auto: input.auto || input.yolo,
       }),
     )
   }),

packages/cli/src/node/target.ts

@@ -11,8 +11,6 @@ export function nodeTarget(platform: string, arch: string) {
   const targetArch = arch as "arm64" | "x64"
   const nodePtyPackage = `@lydell/node-pty-${targetPlatform}-${targetArch}`
   const parcelWatcherPackage = `@parcel/watcher-${targetPlatform}-${targetArch}${targetPlatform === "linux" ? "-glibc" : ""}`
-  const fffPackage = `@ff-labs/fff-bin-${targetPlatform}-${targetArch}${targetPlatform === "linux" ? "-gnu" : ""}`
-  const fffFfiPackage = `@yuuang/ffi-rs-${targetPlatform}-${targetArch}${targetPlatform === "linux" ? "-gnu" : targetPlatform === "win32" ? "-msvc" : ""}`
 
   return {
     platform: targetPlatform,
@@ -21,10 +19,6 @@ export function nodeTarget(platform: string, arch: string) {
     nodePtyEntryAsset: `${nodePtyPackage}/lib/index.js`,
     parcelWatcherPackage,
     parcelWatcherAsset: `${parcelWatcherPackage}/watcher.node`,
-    fffPackage,
-    fffAsset: `${fffPackage}/${targetPlatform === "darwin" ? "libfff_c.dylib" : targetPlatform === "win32" ? "fff_c.dll" : "libfff_c.so"}`,
-    fffFfiPackage,
-    fffFfiAsset: `${fffFfiPackage}/ffi-rs.${targetPlatform}-${targetArch}${targetPlatform === "linux" ? "-gnu" : targetPlatform === "win32" ? "-msvc" : ""}.node`,
   }
 }
 

packages/cli/src/session-target.ts

@@ -105,7 +105,7 @@ async function selectSession(input: {
     return {
       session: input.fork
         ? await input.client.session
-            .fork({ sessionID: explicit.id, boundary: { type: "through" } }, ...requestOptions(input.signal))
+            .fork({ sessionID: explicit.id }, ...requestOptions(input.signal))
             .catch((error) => {
               throw new SessionTargetMutationError(error)
             })
@@ -118,11 +118,9 @@ async function selectSession(input: {
   if (!selected) return { session: undefined, location }
   return {
     session: input.fork
-      ? await input.client.session
-          .fork({ sessionID: selected.id, boundary: { type: "through" } }, ...requestOptions(input.signal))
-          .catch((error) => {
-            throw new SessionTargetMutationError(error)
-          })
+      ? await input.client.session.fork({ sessionID: selected.id }, ...requestOptions(input.signal)).catch((error) => {
+          throw new SessionTargetMutationError(error)
+        })
       : selected,
   }
 }

packages/cli/test/run/noninteractive.test.ts

@@ -267,7 +267,7 @@ async function run(input: {
     values.push(...input.turn(messageID))
     wake?.()
     wake = undefined
-    return ok({ id: messageID, sessionID: "ses_1", timeCreated: 1 }) as never
+    return ok({ admittedSeq: 1, id: messageID, sessionID: "ses_1", timeCreated: 1 }) as never
   })
   await runNonInteractivePrompt({
     client: sdk,

packages/cli/vite.node.config.ts

@@ -30,44 +30,6 @@ function runtimeRequirePlugin(): Plugin {
   }
 }
 
-function fffNodePlugin(): Plugin {
-  return {
-    name: "opencode:fff-node",
-    enforce: "pre",
-    transform(code, id) {
-      const normalized = id.replaceAll("\\", "/")
-      if (normalized.endsWith("/ffi-rs/index.js")) {
-        const start = code.indexOf("if (!nativeBinding) {")
-        if (start === -1) this.error("Failed to rewrite ffi-rs native binding loader")
-        return `const unavailable = () => { throw new Error("ffi-rs native binding unavailable") }
-const nativeBinding = globalThis.__OPENCODE_FFF_FFI ?? {
-  DataType: new Proxy({}, { get: (target, key) => target[key] ?? key }),
-  PointerType: {},
-  FFITypeTag: {},
-  open: unavailable,
-  close: unavailable,
-  load: unavailable,
-  isNullPointer: unavailable,
-  createPointer: unavailable,
-  restorePointer: unavailable,
-  unwrapPointer: unavailable,
-  wrapPointer: unavailable,
-  freePointer: unavailable,
-}
-const loadError = undefined
-${code.slice(start)}`
-      }
-      if (!normalized.endsWith("/fff-node/dist/src/binary.js")) return
-      const transformed = code.replace(
-        "export function findBinary() {",
-        "export function findBinary() { if (process.env.FFF_BINARY_PATH) return process.env.FFF_BINARY_PATH;",
-      )
-      if (transformed === code) this.error("Failed to rewrite FFF binary loader")
-      return transformed
-    },
-  }
-}
-
 const resolve = {
   alias: [
     { find: /^solid-js\/store$/, replacement: "solid-js/store/dist/store.js" },
@@ -194,11 +156,6 @@ process.env.OTUI_ASSET_ROOT = __ocAssetRoot
 process.env.OPENCODE_NODE_PTY_PATH = __ocPath.join(__ocAssetRoot, ${JSON.stringify(input.target.nodePtyEntryAsset)})
 process.env.OPENCODE_PARCEL_WATCHER_PATH = __ocPath.join(__ocAssetRoot, ${JSON.stringify(input.target.parcelWatcherAsset)})
 process.env.OPENCODE_PHOTON_WASM_PATH = __ocPath.join(__ocAssetRoot, ${JSON.stringify(photonWasmAsset)})
-process.env.FFF_BINARY_PATH = __ocPath.join(__ocAssetRoot, ${JSON.stringify(input.target.fffAsset)})
-process.env.OPENCODE_FFF_FFI_PATH = __ocPath.join(__ocAssetRoot, ${JSON.stringify(input.target.fffFfiAsset)})
-try {
-  globalThis.__OPENCODE_FFF_FFI = require(process.env.OPENCODE_FFF_FFI_PATH)
-} catch {}
 globalThis.__OPENCODE_PHOTON_WASM_PATH = process.env.OPENCODE_PHOTON_WASM_PATH
 if (process.platform === "linux") process.env.OPENTUI_LIBC = "glibc"`
 }
@@ -217,7 +174,6 @@ export function mainConfig(input: NodeBuildInput): UserConfig {
     plugins: [
       rawTextPlugin(),
       runtimeRequirePlugin(),
-      fffNodePlugin(),
       solid({
         solid: {
           generate: "universal",

packages/client/src/effect/api/api.ts

@@ -136,7 +136,7 @@ export type Endpoint5_4Input = { readonly sessionID: Session.ID }
 export type Endpoint5_4Output = void
 export type SessionRemoveOperation<E = never> = (input: Endpoint5_4Input) => Effect.Effect<Endpoint5_4Output, E>
 
-export type Endpoint5_5Input = { readonly sessionID: Session.ID; readonly boundary: Session.ForkRequestBoundary }
+export type Endpoint5_5Input = { readonly sessionID: Session.ID; readonly messageID?: SessionMessage.ID | undefined }
 export type Endpoint5_5Output = Session.Info
 export type SessionForkOperation<E = never> = (input: Endpoint5_5Input) => Effect.Effect<Endpoint5_5Output, E>
 
@@ -152,15 +152,19 @@ export type Endpoint5_8Input = { readonly sessionID: Session.ID; readonly title:
 export type Endpoint5_8Output = void
 export type SessionRenameOperation<E = never> = (input: Endpoint5_8Input) => Effect.Effect<Endpoint5_8Output, E>
 
-export type Endpoint5_9Input = {
+export type Endpoint5_9Input = { readonly sessionID: Session.ID }
+export type Endpoint5_9Output = void
+export type SessionArchiveOperation<E = never> = (input: Endpoint5_9Input) => Effect.Effect<Endpoint5_9Output, E>
+
+export type Endpoint5_10Input = {
   readonly sessionID: Session.ID
   readonly directory: AbsolutePath
   readonly workspaceID?: Workspace.ID | undefined
 }
-export type Endpoint5_9Output = void
-export type SessionMoveOperation<E = never> = (input: Endpoint5_9Input) => Effect.Effect<Endpoint5_9Output, E>
+export type Endpoint5_10Output = void
+export type SessionMoveOperation<E = never> = (input: Endpoint5_10Input) => Effect.Effect<Endpoint5_10Output, E>
 
-export type Endpoint5_10Input = {
+export type Endpoint5_11Input = {
   readonly sessionID: Session.ID
   readonly id?: SessionMessage.ID | undefined
   readonly text: string
@@ -170,10 +174,10 @@ export type Endpoint5_10Input = {
   readonly delivery?: "steer" | "queue" | undefined
   readonly resume?: boolean | undefined
 }
-export type Endpoint5_10Output = SessionPending.User
-export type SessionPromptOperation<E = never> = (input: Endpoint5_10Input) => Effect.Effect<Endpoint5_10Output, E>
+export type Endpoint5_11Output = SessionPending.User
+export type SessionPromptOperation<E = never> = (input: Endpoint5_11Input) => Effect.Effect<Endpoint5_11Output, E>
 
-export type Endpoint5_11Input = {
+export type Endpoint5_12Input = {
   readonly sessionID: Session.ID
   readonly id?: SessionMessage.ID | undefined
   readonly command: string
@@ -185,19 +189,19 @@ export type Endpoint5_11Input = {
   readonly delivery?: "steer" | "queue" | undefined
   readonly resume?: boolean | undefined
 }
-export type Endpoint5_11Output = SessionPending.User
-export type SessionCommandOperation<E = never> = (input: Endpoint5_11Input) => Effect.Effect<Endpoint5_11Output, E>
+export type Endpoint5_12Output = SessionPending.User
+export type SessionCommandOperation<E = never> = (input: Endpoint5_12Input) => Effect.Effect<Endpoint5_12Output, E>
 
-export type Endpoint5_12Input = {
+export type Endpoint5_13Input = {
   readonly sessionID: Session.ID
   readonly id?: SessionMessage.ID | undefined
   readonly skill: Skill.ID
   readonly resume?: boolean | undefined
 }
-export type Endpoint5_12Output = void
-export type SessionSkillOperation<E = never> = (input: Endpoint5_12Input) => Effect.Effect<Endpoint5_12Output, E>
+export type Endpoint5_13Output = void
+export type SessionSkillOperation<E = never> = (input: Endpoint5_13Input) => Effect.Effect<Endpoint5_13Output, E>
 
-export type Endpoint5_13Input = {
+export type Endpoint5_14Input = {
   readonly sessionID: Session.ID
   readonly id?: SessionMessage.ID | undefined
   readonly text: string
@@ -206,81 +210,81 @@ export type Endpoint5_13Input = {
   readonly delivery?: "steer" | "queue" | undefined
   readonly resume?: boolean | undefined
 }
-export type Endpoint5_13Output = SessionPending.Synthetic
-export type SessionSyntheticOperation<E = never> = (input: Endpoint5_13Input) => Effect.Effect<Endpoint5_13Output, E>
+export type Endpoint5_14Output = SessionPending.Synthetic
+export type SessionSyntheticOperation<E = never> = (input: Endpoint5_14Input) => Effect.Effect<Endpoint5_14Output, E>
 
-export type Endpoint5_14Input = {
+export type Endpoint5_15Input = {
   readonly sessionID: Session.ID
   readonly id?: Event.ID | undefined
   readonly command: string
 }
-export type Endpoint5_14Output = void
-export type SessionShellOperation<E = never> = (input: Endpoint5_14Input) => Effect.Effect<Endpoint5_14Output, E>
+export type Endpoint5_15Output = void
+export type SessionShellOperation<E = never> = (input: Endpoint5_15Input) => Effect.Effect<Endpoint5_15Output, E>
 
-export type Endpoint5_15Input = { readonly sessionID: Session.ID; readonly id?: SessionMessage.ID | undefined }
-export type Endpoint5_15Output = SessionPending.Compaction
-export type SessionCompactOperation<E = never> = (input: Endpoint5_15Input) => Effect.Effect<Endpoint5_15Output, E>
+export type Endpoint5_16Input = { readonly sessionID: Session.ID; readonly id?: SessionMessage.ID | undefined }
+export type Endpoint5_16Output = SessionPending.Compaction
+export type SessionCompactOperation<E = never> = (input: Endpoint5_16Input) => Effect.Effect<Endpoint5_16Output, E>
 
-export type Endpoint5_16Input = { readonly sessionID: Session.ID }
-export type Endpoint5_16Output = void
-export type SessionWaitOperation<E = never> = (input: Endpoint5_16Input) => Effect.Effect<Endpoint5_16Output, E>
+export type Endpoint5_17Input = { readonly sessionID: Session.ID }
+export type Endpoint5_17Output = void
+export type SessionWaitOperation<E = never> = (input: Endpoint5_17Input) => Effect.Effect<Endpoint5_17Output, E>
 
-export type Endpoint5_17Input = {
+export type Endpoint5_18Input = {
   readonly sessionID: Session.ID
   readonly messageID: SessionMessage.ID
   readonly files?: boolean | undefined
 }
-export type Endpoint5_17Output = Session.Revert
-export type SessionRevertStageOperation<E = never> = (input: Endpoint5_17Input) => Effect.Effect<Endpoint5_17Output, E>
-
-export type Endpoint5_18Input = { readonly sessionID: Session.ID }
-export type Endpoint5_18Output = void
-export type SessionRevertClearOperation<E = never> = (input: Endpoint5_18Input) => Effect.Effect<Endpoint5_18Output, E>
+export type Endpoint5_18Output = Session.Revert
+export type SessionRevertStageOperation<E = never> = (input: Endpoint5_18Input) => Effect.Effect<Endpoint5_18Output, E>
 
 export type Endpoint5_19Input = { readonly sessionID: Session.ID }
 export type Endpoint5_19Output = void
-export type SessionRevertCommitOperation<E = never> = (input: Endpoint5_19Input) => Effect.Effect<Endpoint5_19Output, E>
+export type SessionRevertClearOperation<E = never> = (input: Endpoint5_19Input) => Effect.Effect<Endpoint5_19Output, E>
 
 export type Endpoint5_20Input = { readonly sessionID: Session.ID }
-export type Endpoint5_20Output = ReadonlyArray<SessionMessage.Info>
-export type SessionContextOperation<E = never> = (input: Endpoint5_20Input) => Effect.Effect<Endpoint5_20Output, E>
+export type Endpoint5_20Output = void
+export type SessionRevertCommitOperation<E = never> = (input: Endpoint5_20Input) => Effect.Effect<Endpoint5_20Output, E>
 
 export type Endpoint5_21Input = { readonly sessionID: Session.ID }
-export type Endpoint5_21Output = ReadonlyArray<SessionPending.Info>
-export type SessionPendingListOperation<E = never> = (input: Endpoint5_21Input) => Effect.Effect<Endpoint5_21Output, E>
+export type Endpoint5_21Output = ReadonlyArray<SessionMessage.Info>
+export type SessionContextOperation<E = never> = (input: Endpoint5_21Input) => Effect.Effect<Endpoint5_21Output, E>
 
 export type Endpoint5_22Input = { readonly sessionID: Session.ID }
-export type Endpoint5_22Output = ReadonlyArray<InstructionEntry.Info>
-export type SessionInstructionsEntryListOperation<E = never> = (
-  input: Endpoint5_22Input,
-) => Effect.Effect<Endpoint5_22Output, E>
+export type Endpoint5_22Output = ReadonlyArray<SessionPending.Info>
+export type SessionPendingListOperation<E = never> = (input: Endpoint5_22Input) => Effect.Effect<Endpoint5_22Output, E>
 
-export type Endpoint5_23Input = {
+export type Endpoint5_23Input = { readonly sessionID: Session.ID }
+export type Endpoint5_23Output = ReadonlyArray<InstructionEntry.Info>
+export type SessionInstructionsEntryListOperation<E = never> = (
+  input: Endpoint5_23Input,
+) => Effect.Effect<Endpoint5_23Output, E>
+
+export type Endpoint5_24Input = {
   readonly sessionID: Session.ID
   readonly key: InstructionEntry.Key
   readonly value: Schema.Json
 }
-export type Endpoint5_23Output = void
-export type SessionInstructionsEntryPutOperation<E = never> = (
-  input: Endpoint5_23Input,
-) => Effect.Effect<Endpoint5_23Output, E>
-
-export type Endpoint5_24Input = { readonly sessionID: Session.ID; readonly key: InstructionEntry.Key }
 export type Endpoint5_24Output = void
-export type SessionInstructionsEntryRemoveOperation<E = never> = (
+export type SessionInstructionsEntryPutOperation<E = never> = (
   input: Endpoint5_24Input,
 ) => Effect.Effect<Endpoint5_24Output, E>
 
-export type Endpoint5_25Input = { readonly sessionID: Session.ID; readonly prompt: string }
-export type Endpoint5_25Output = { readonly text: string }
-export type SessionGenerateOperation<E = never> = (input: Endpoint5_25Input) => Effect.Effect<Endpoint5_25Output, E>
+export type Endpoint5_25Input = { readonly sessionID: Session.ID; readonly key: InstructionEntry.Key }
+export type Endpoint5_25Output = void
+export type SessionInstructionsEntryRemoveOperation<E = never> = (
+  input: Endpoint5_25Input,
+) => Effect.Effect<Endpoint5_25Output, E>
 
-export type Endpoint5_26Input = {
+export type Endpoint5_26Input = { readonly sessionID: Session.ID; readonly prompt: string }
+export type Endpoint5_26Output = { readonly text: string }
+export type SessionGenerateOperation<E = never> = (input: Endpoint5_26Input) => Effect.Effect<Endpoint5_26Output, E>
+
+export type Endpoint5_27Input = {
   readonly sessionID: Session.ID
   readonly after?: Event.Seq | undefined
   readonly follow?: boolean | undefined
 }
-export type Endpoint5_26Output =
+export type Endpoint5_27Output =
   | (
       | {
           readonly id: Event.ID
@@ -323,6 +327,15 @@ export type Endpoint5_26Output =
           readonly location?: Location.Ref | undefined
           readonly data: { readonly sessionID: Session.ID; readonly title: string }
         }
+      | {
+          readonly id: Event.ID
+          readonly created: DateTime.Utc
+          readonly metadata?: { readonly [x: string]: unknown } | undefined
+          readonly type: "session.archived"
+          readonly durable: { readonly aggregateID: string; readonly seq: Event.Seq; readonly version: Event.Version }
+          readonly location?: Location.Ref | undefined
+          readonly data: { readonly sessionID: Session.ID }
+        }
       | {
           readonly id: Event.ID
           readonly created: DateTime.Utc
@@ -342,10 +355,8 @@ export type Endpoint5_26Output =
           readonly data: {
             readonly sessionID: Session.ID
             readonly parentID: Session.ID
-            readonly boundary: Session.ForkBoundary
-            readonly instructions?:
-              | { readonly [x: string & Brand.Brand<"Instruction.Key">]: string & Brand.Brand<"Instruction.Hash"> }
-              | undefined
+            readonly parentSeq: number
+            readonly from?: SessionMessage.ID | undefined
           }
         }
       | {
@@ -827,19 +838,19 @@ export type Endpoint5_26Output =
         }
     )
   | EventLog.Synced
-export type SessionLogOperation<E = never> = (input: Endpoint5_26Input) => Stream.Stream<Endpoint5_26Output, E>
-
-export type Endpoint5_27Input = { readonly sessionID: Session.ID }
-export type Endpoint5_27Output = void
-export type SessionInterruptOperation<E = never> = (input: Endpoint5_27Input) => Effect.Effect<Endpoint5_27Output, E>
+export type SessionLogOperation<E = never> = (input: Endpoint5_27Input) => Stream.Stream<Endpoint5_27Output, E>
 
 export type Endpoint5_28Input = { readonly sessionID: Session.ID }
 export type Endpoint5_28Output = void
-export type SessionBackgroundOperation<E = never> = (input: Endpoint5_28Input) => Effect.Effect<Endpoint5_28Output, E>
+export type SessionInterruptOperation<E = never> = (input: Endpoint5_28Input) => Effect.Effect<Endpoint5_28Output, E>
 
-export type Endpoint5_29Input = { readonly sessionID: Session.ID; readonly messageID: SessionMessage.ID }
-export type Endpoint5_29Output = SessionMessage.Info
-export type SessionMessageOperation<E = never> = (input: Endpoint5_29Input) => Effect.Effect<Endpoint5_29Output, E>
+export type Endpoint5_29Input = { readonly sessionID: Session.ID }
+export type Endpoint5_29Output = void
+export type SessionBackgroundOperation<E = never> = (input: Endpoint5_29Input) => Effect.Effect<Endpoint5_29Output, E>
+
+export type Endpoint5_30Input = { readonly sessionID: Session.ID; readonly messageID: SessionMessage.ID }
+export type Endpoint5_30Output = SessionMessage.Info
+export type SessionMessageOperation<E = never> = (input: Endpoint5_30Input) => Effect.Effect<Endpoint5_30Output, E>
 
 export interface SessionApi<E = never> {
   readonly list: SessionListOperation<E>
@@ -851,6 +862,7 @@ export interface SessionApi<E = never> {
   readonly switchAgent: SessionSwitchAgentOperation<E>
   readonly switchModel: SessionSwitchModelOperation<E>
   readonly rename: SessionRenameOperation<E>
+  readonly archive: SessionArchiveOperation<E>
   readonly move: SessionMoveOperation<E>
   readonly prompt: SessionPromptOperation<E>
   readonly command: SessionCommandOperation<E>

packages/client/src/effect/generated/client.ts

@@ -76,6 +76,8 @@ import type {
   Endpoint5_28Output,
   Endpoint5_29Input,
   Endpoint5_29Output,
+  Endpoint5_30Input,
+  Endpoint5_30Output,
   Endpoint6_0Input,
   Endpoint6_0Output,
   Endpoint7_0Input,
@@ -329,7 +331,7 @@ const Endpoint5_4 = (raw: RawClient["server.session"]) => (input: Endpoint5_4Inp
 
 const Endpoint5_5 = (raw: RawClient["server.session"]) => (input: Endpoint5_5Input) =>
   preserveEffect<Endpoint5_5Output>()(
-    raw["session.fork"]({ params: { sessionID: input["sessionID"] }, payload: { boundary: input["boundary"] } }).pipe(
+    raw["session.fork"]({ params: { sessionID: input["sessionID"] }, payload: { messageID: input["messageID"] } }).pipe(
       Effect.mapError(mapClientError),
       Effect.map((value) => value.data),
     ),
@@ -358,14 +360,19 @@ const Endpoint5_8 = (raw: RawClient["server.session"]) => (input: Endpoint5_8Inp
 
 const Endpoint5_9 = (raw: RawClient["server.session"]) => (input: Endpoint5_9Input) =>
   preserveEffect<Endpoint5_9Output>()(
+    raw["session.archive"]({ params: { sessionID: input["sessionID"] } }).pipe(Effect.mapError(mapClientError)),
+  )
+
+const Endpoint5_10 = (raw: RawClient["server.session"]) => (input: Endpoint5_10Input) =>
+  preserveEffect<Endpoint5_10Output>()(
     raw["session.move"]({
       params: { sessionID: input["sessionID"] },
       payload: { directory: input["directory"], workspaceID: input["workspaceID"] },
     }).pipe(Effect.mapError(mapClientError)),
   )
 
-const Endpoint5_10 = (raw: RawClient["server.session"]) => (input: Endpoint5_10Input) =>
-  preserveEffect<Endpoint5_10Output>()(
+const Endpoint5_11 = (raw: RawClient["server.session"]) => (input: Endpoint5_11Input) =>
+  preserveEffect<Endpoint5_11Output>()(
     raw["session.prompt"]({
       params: { sessionID: input["sessionID"] },
       payload: {
@@ -383,8 +390,8 @@ const Endpoint5_10 = (raw: RawClient["server.session"]) => (input: Endpoint5_10I
     ),
   )
 
-const Endpoint5_11 = (raw: RawClient["server.session"]) => (input: Endpoint5_11Input) =>
-  preserveEffect<Endpoint5_11Output>()(
+const Endpoint5_12 = (raw: RawClient["server.session"]) => (input: Endpoint5_12Input) =>
+  preserveEffect<Endpoint5_12Output>()(
     raw["session.command"]({
       params: { sessionID: input["sessionID"] },
       payload: {
@@ -404,16 +411,16 @@ const Endpoint5_11 = (raw: RawClient["server.session"]) => (input: Endpoint5_11I
     ),
   )
 
-const Endpoint5_12 = (raw: RawClient["server.session"]) => (input: Endpoint5_12Input) =>
-  preserveEffect<Endpoint5_12Output>()(
+const Endpoint5_13 = (raw: RawClient["server.session"]) => (input: Endpoint5_13Input) =>
+  preserveEffect<Endpoint5_13Output>()(
     raw["session.skill"]({
       params: { sessionID: input["sessionID"] },
       payload: { id: input["id"], skill: input["skill"], resume: input["resume"] },
     }).pipe(Effect.mapError(mapClientError)),
   )
 
-const Endpoint5_13 = (raw: RawClient["server.session"]) => (input: Endpoint5_13Input) =>
-  preserveEffect<Endpoint5_13Output>()(
+const Endpoint5_14 = (raw: RawClient["server.session"]) => (input: Endpoint5_14Input) =>
+  preserveEffect<Endpoint5_14Output>()(
     raw["session.synthetic"]({
       params: { sessionID: input["sessionID"] },
       payload: {
@@ -430,29 +437,29 @@ const Endpoint5_13 = (raw: RawClient["server.session"]) => (input: Endpoint5_13I
     ),
   )
 
-const Endpoint5_14 = (raw: RawClient["server.session"]) => (input: Endpoint5_14Input) =>
-  preserveEffect<Endpoint5_14Output>()(
+const Endpoint5_15 = (raw: RawClient["server.session"]) => (input: Endpoint5_15Input) =>
+  preserveEffect<Endpoint5_15Output>()(
     raw["session.shell"]({
       params: { sessionID: input["sessionID"] },
       payload: { id: input["id"], command: input["command"] },
     }).pipe(Effect.mapError(mapClientError)),
   )
 
-const Endpoint5_15 = (raw: RawClient["server.session"]) => (input: Endpoint5_15Input) =>
-  preserveEffect<Endpoint5_15Output>()(
+const Endpoint5_16 = (raw: RawClient["server.session"]) => (input: Endpoint5_16Input) =>
+  preserveEffect<Endpoint5_16Output>()(
     raw["session.compact"]({ params: { sessionID: input["sessionID"] }, payload: { id: input["id"] } }).pipe(
       Effect.mapError(mapClientError),
       Effect.map((value) => value.data),
     ),
   )
 
-const Endpoint5_16 = (raw: RawClient["server.session"]) => (input: Endpoint5_16Input) =>
-  preserveEffect<Endpoint5_16Output>()(
+const Endpoint5_17 = (raw: RawClient["server.session"]) => (input: Endpoint5_17Input) =>
+  preserveEffect<Endpoint5_17Output>()(
     raw["session.wait"]({ params: { sessionID: input["sessionID"] } }).pipe(Effect.mapError(mapClientError)),
   )
 
-const Endpoint5_17 = (raw: RawClient["server.session"]) => (input: Endpoint5_17Input) =>
-  preserveEffect<Endpoint5_17Output>()(
+const Endpoint5_18 = (raw: RawClient["server.session"]) => (input: Endpoint5_18Input) =>
+  preserveEffect<Endpoint5_18Output>()(
     raw["session.revert.stage"]({
       params: { sessionID: input["sessionID"] },
       payload: { messageID: input["messageID"], files: input["files"] },
@@ -462,27 +469,19 @@ const Endpoint5_17 = (raw: RawClient["server.session"]) => (input: Endpoint5_17I
     ),
   )
 
-const Endpoint5_18 = (raw: RawClient["server.session"]) => (input: Endpoint5_18Input) =>
-  preserveEffect<Endpoint5_18Output>()(
-    raw["session.revert.clear"]({ params: { sessionID: input["sessionID"] } }).pipe(Effect.mapError(mapClientError)),
-  )
-
 const Endpoint5_19 = (raw: RawClient["server.session"]) => (input: Endpoint5_19Input) =>
   preserveEffect<Endpoint5_19Output>()(
-    raw["session.revert.commit"]({ params: { sessionID: input["sessionID"] } }).pipe(Effect.mapError(mapClientError)),
+    raw["session.revert.clear"]({ params: { sessionID: input["sessionID"] } }).pipe(Effect.mapError(mapClientError)),
   )
 
 const Endpoint5_20 = (raw: RawClient["server.session"]) => (input: Endpoint5_20Input) =>
   preserveEffect<Endpoint5_20Output>()(
-    raw["session.context"]({ params: { sessionID: input["sessionID"] } }).pipe(
-      Effect.mapError(mapClientError),
-      Effect.map((value) => value.data),
-    ),
+    raw["session.revert.commit"]({ params: { sessionID: input["sessionID"] } }).pipe(Effect.mapError(mapClientError)),
   )
 
 const Endpoint5_21 = (raw: RawClient["server.session"]) => (input: Endpoint5_21Input) =>
   preserveEffect<Endpoint5_21Output>()(
-    raw["session.pending.list"]({ params: { sessionID: input["sessionID"] } }).pipe(
+    raw["session.context"]({ params: { sessionID: input["sessionID"] } }).pipe(
       Effect.mapError(mapClientError),
       Effect.map((value) => value.data),
     ),
@@ -490,7 +489,7 @@ const Endpoint5_21 = (raw: RawClient["server.session"]) => (input: Endpoint5_21I
 
 const Endpoint5_22 = (raw: RawClient["server.session"]) => (input: Endpoint5_22Input) =>
   preserveEffect<Endpoint5_22Output>()(
-    raw["session.instructions.entry.list"]({ params: { sessionID: input["sessionID"] } }).pipe(
+    raw["session.pending.list"]({ params: { sessionID: input["sessionID"] } }).pipe(
       Effect.mapError(mapClientError),
       Effect.map((value) => value.data),
     ),
@@ -498,29 +497,37 @@ const Endpoint5_22 = (raw: RawClient["server.session"]) => (input: Endpoint5_22I
 
 const Endpoint5_23 = (raw: RawClient["server.session"]) => (input: Endpoint5_23Input) =>
   preserveEffect<Endpoint5_23Output>()(
+    raw["session.instructions.entry.list"]({ params: { sessionID: input["sessionID"] } }).pipe(
+      Effect.mapError(mapClientError),
+      Effect.map((value) => value.data),
+    ),
+  )
+
+const Endpoint5_24 = (raw: RawClient["server.session"]) => (input: Endpoint5_24Input) =>
+  preserveEffect<Endpoint5_24Output>()(
     raw["session.instructions.entry.put"]({
       params: { sessionID: input["sessionID"], key: input["key"] },
       payload: { value: input["value"] },
     }).pipe(Effect.mapError(mapClientError)),
   )
 
-const Endpoint5_24 = (raw: RawClient["server.session"]) => (input: Endpoint5_24Input) =>
-  preserveEffect<Endpoint5_24Output>()(
+const Endpoint5_25 = (raw: RawClient["server.session"]) => (input: Endpoint5_25Input) =>
+  preserveEffect<Endpoint5_25Output>()(
     raw["session.instructions.entry.remove"]({ params: { sessionID: input["sessionID"], key: input["key"] } }).pipe(
       Effect.mapError(mapClientError),
     ),
   )
 
-const Endpoint5_25 = (raw: RawClient["server.session"]) => (input: Endpoint5_25Input) =>
-  preserveEffect<Endpoint5_25Output>()(
+const Endpoint5_26 = (raw: RawClient["server.session"]) => (input: Endpoint5_26Input) =>
+  preserveEffect<Endpoint5_26Output>()(
     raw["session.generate"]({ params: { sessionID: input["sessionID"] }, payload: { prompt: input["prompt"] } }).pipe(
       Effect.mapError(mapClientError),
       Effect.map((value) => value.data),
     ),
   )
 
-const Endpoint5_26 = (raw: RawClient["server.session"]) => (input: Endpoint5_26Input) =>
-  preserveStream<Endpoint5_26Output>()(
+const Endpoint5_27 = (raw: RawClient["server.session"]) => (input: Endpoint5_27Input) =>
+  preserveStream<Endpoint5_27Output>()(
     Stream.unwrap(
       raw["session.log"]({
         params: { sessionID: input["sessionID"] },
@@ -532,18 +539,18 @@ const Endpoint5_26 = (raw: RawClient["server.session"]) => (input: Endpoint5_26I
     ),
   )
 
-const Endpoint5_27 = (raw: RawClient["server.session"]) => (input: Endpoint5_27Input) =>
-  preserveEffect<Endpoint5_27Output>()(
-    raw["session.interrupt"]({ params: { sessionID: input["sessionID"] } }).pipe(Effect.mapError(mapClientError)),
-  )
-
 const Endpoint5_28 = (raw: RawClient["server.session"]) => (input: Endpoint5_28Input) =>
   preserveEffect<Endpoint5_28Output>()(
-    raw["session.background"]({ params: { sessionID: input["sessionID"] } }).pipe(Effect.mapError(mapClientError)),
+    raw["session.interrupt"]({ params: { sessionID: input["sessionID"] } }).pipe(Effect.mapError(mapClientError)),
   )
 
 const Endpoint5_29 = (raw: RawClient["server.session"]) => (input: Endpoint5_29Input) =>
   preserveEffect<Endpoint5_29Output>()(
+    raw["session.background"]({ params: { sessionID: input["sessionID"] } }).pipe(Effect.mapError(mapClientError)),
+  )
+
+const Endpoint5_30 = (raw: RawClient["server.session"]) => (input: Endpoint5_30Input) =>
+  preserveEffect<Endpoint5_30Output>()(
     raw["session.message"]({ params: { sessionID: input["sessionID"], messageID: input["messageID"] } }).pipe(
       Effect.mapError(mapClientError),
       Effect.map((value) => value.data),
@@ -560,23 +567,24 @@ const adaptGroup5 = (raw: RawClient["server.session"]) => ({
   switchAgent: Endpoint5_6(raw),
   switchModel: Endpoint5_7(raw),
   rename: Endpoint5_8(raw),
-  move: Endpoint5_9(raw),
-  prompt: Endpoint5_10(raw),
-  command: Endpoint5_11(raw),
-  skill: Endpoint5_12(raw),
-  synthetic: Endpoint5_13(raw),
-  shell: Endpoint5_14(raw),
-  compact: Endpoint5_15(raw),
-  wait: Endpoint5_16(raw),
-  revert: { stage: Endpoint5_17(raw), clear: Endpoint5_18(raw), commit: Endpoint5_19(raw) },
-  context: Endpoint5_20(raw),
-  pending: { list: Endpoint5_21(raw) },
-  instructions: { entry: { list: Endpoint5_22(raw), put: Endpoint5_23(raw), remove: Endpoint5_24(raw) } },
-  generate: Endpoint5_25(raw),
-  log: Endpoint5_26(raw),
-  interrupt: Endpoint5_27(raw),
-  background: Endpoint5_28(raw),
-  message: Endpoint5_29(raw),
+  archive: Endpoint5_9(raw),
+  move: Endpoint5_10(raw),
+  prompt: Endpoint5_11(raw),
+  command: Endpoint5_12(raw),
+  skill: Endpoint5_13(raw),
+  synthetic: Endpoint5_14(raw),
+  shell: Endpoint5_15(raw),
+  compact: Endpoint5_16(raw),
+  wait: Endpoint5_17(raw),
+  revert: { stage: Endpoint5_18(raw), clear: Endpoint5_19(raw), commit: Endpoint5_20(raw) },
+  context: Endpoint5_21(raw),
+  pending: { list: Endpoint5_22(raw) },
+  instructions: { entry: { list: Endpoint5_23(raw), put: Endpoint5_24(raw), remove: Endpoint5_25(raw) } },
+  generate: Endpoint5_26(raw),
+  log: Endpoint5_27(raw),
+  interrupt: Endpoint5_28(raw),
+  background: Endpoint5_29(raw),
+  message: Endpoint5_30(raw),
 })
 
 const Endpoint6_0 = (raw: RawClient["server.message"]) => (input: Endpoint6_0Input) =>

packages/client/src/promise/generated/client.ts

@@ -28,6 +28,8 @@ import type {
   SessionSwitchModelOutput,
   SessionRenameInput,
   SessionRenameOutput,
+  SessionArchiveInput,
+  SessionArchiveOutput,
   SessionMoveInput,
   SessionMoveOutput,
   SessionPromptInput,
@@ -510,7 +512,7 @@ export function make(options: ClientOptions) {
           {
             method: "POST",
             path: `/api/session/${encodeURIComponent(input.sessionID)}/fork`,
-            body: { boundary: input["boundary"] },
+            body: { messageID: input["messageID"] },
             successStatus: 200,
             declaredStatuses: [404, 400, 401],
             empty: false,
@@ -553,6 +555,17 @@ export function make(options: ClientOptions) {
           },
           requestOptions,
         ),
+      archive: (input: SessionArchiveInput, requestOptions?: RequestOptions) =>
+        request<SessionArchiveOutput>(
+          {
+            method: "POST",
+            path: `/api/session/${encodeURIComponent(input.sessionID)}/archive`,
+            successStatus: 204,
+            declaredStatuses: [404, 400, 401],
+            empty: true,
+          },
+          requestOptions,
+        ),
       move: (input: SessionMoveInput, requestOptions?: RequestOptions) =>
         request<SessionMoveOutput>(
           {

packages/client/src/promise/generated/types.ts

@@ -14,8 +14,6 @@ export type PermissionEffect = "allow" | "deny" | "ask"
 
 export type PluginInfo = { id: string }
 
-export type SessionForkBoundary = { type: "before"; messageID: string } | { type: "through"; messageID: string }
-
 export type MoneyUSD = number
 
 export type TokenUsageInfo = {
@@ -45,7 +43,13 @@ export type PromptMention = { start: number; end: number; text: string }
 
 export type SessionPendingSyntheticData = { text: string; description?: string; metadata?: { [x: string]: JsonValue } }
 
-export type SessionPendingCompaction = { id: string; sessionID: string; timeCreated: number; type: "compaction" }
+export type SessionPendingCompaction = {
+  admittedSeq: number
+  id: string
+  sessionID: string
+  timeCreated: number
+  type: "compaction"
+}
 
 export type SessionMessageAgentSelected = {
   id: string
@@ -607,6 +611,16 @@ export type SessionRenamed = {
   data: { sessionID: string; title: string }
 }
 
+export type SessionArchived = {
+  id: string
+  created: number
+  metadata?: { [x: string]: any }
+  type: "session.archived"
+  durable: { aggregateID: string; seq: number; version: 1 }
+  location?: LocationRef
+  data: { sessionID: string }
+}
+
 export type SessionDeleted = {
   id: string
   created: number
@@ -624,7 +638,7 @@ export type SessionForked = {
   type: "session.forked"
   durable: { aggregateID: string; seq: number; version: 2 }
   location?: LocationRef
-  data: { sessionID: string; parentID: string; boundary: SessionForkBoundary; instructions?: { [x: string]: string } }
+  data: { sessionID: string; parentID: string; parentSeq: number; from?: string }
 }
 
 export type SessionInputPromoted = {
@@ -1206,6 +1220,7 @@ export type PromptFileAttachment = {
 export type PromptAgentAttachment = { name: string; mention?: PromptMention }
 
 export type SessionPendingSynthetic = {
+  admittedSeq: number
   id: string
   sessionID: string
   timeCreated: number
@@ -1750,7 +1765,7 @@ export type PermissionRuleset = Array<PermissionRule>
 export type SessionInfo = {
   id: string
   parentID?: string
-  fork?: { sessionID: string; boundary: SessionForkBoundary }
+  fork?: { sessionID: string; messageID?: string }
   projectID: string
   agent?: string
   model?: ModelRef
@@ -1961,6 +1976,7 @@ export type AgentInfo = {
 export type SessionsResponse = { data: Array<SessionInfo>; cursor: { previous?: string | null; next?: string | null } }
 
 export type SessionPendingUser = {
+  admittedSeq: number
   id: string
   sessionID: string
   timeCreated: number
@@ -2153,6 +2169,7 @@ export type SessionEventDurable =
   | SessionModelSelected
   | SessionMoved
   | SessionRenamed
+  | SessionArchived
   | SessionDeleted
   | SessionForked
   | SessionInputPromoted
@@ -2247,6 +2264,7 @@ export type V2Event =
   | SessionModelSelected
   | SessionMoved
   | SessionRenamed
+  | SessionArchived
   | SessionUsageUpdated
   | SessionDeleted
   | SessionForked
@@ -2696,9 +2714,7 @@ export type SessionRemoveOutput = void
 
 export type SessionForkInput = {
   readonly sessionID: { readonly sessionID: string }["sessionID"]
-  readonly boundary: {
-    readonly boundary: { readonly type: "before"; readonly messageID: string } | { readonly type: "through" }
-  }["boundary"]
+  readonly messageID?: { readonly messageID?: string | undefined }["messageID"]
 }
 
 export type SessionForkOutput = { data: SessionInfo }["data"]
@@ -2726,6 +2742,10 @@ export type SessionRenameInput = {
 
 export type SessionRenameOutput = void
 
+export type SessionArchiveInput = { readonly sessionID: { readonly sessionID: string }["sessionID"] }
+
+export type SessionArchiveOutput = void
+
 export type SessionMoveInput = {
   readonly sessionID: { readonly sessionID: string }["sessionID"]
   readonly directory: { readonly directory: string; readonly workspaceID?: string }["directory"]

packages/client/test/effect.test.ts

@@ -268,6 +268,7 @@ const session = {
 
 const admission = {
   data: {
+    admittedSeq: 0,
     id: "msg_test",
     sessionID: "ses_test",
     type: "user",
@@ -280,6 +281,7 @@ const admission = {
 const compactionAdmission = {
   data: {
     type: "compaction",
+    admittedSeq: 1,
     id: "msg_compaction",
     sessionID: "ses_test",
     timeCreated: 1_717_171_717_000,

packages/client/test/promise.test.ts

@@ -304,6 +304,7 @@ test("session.pending.list uses the public HTTP contract", async () => {
   const requests: Array<{ method: string; url: string }> = []
   const pending = [
     {
+      admittedSeq: 3,
       id: "msg_pending",
       sessionID: "ses_test",
       timeCreated: 1_717_171_717_000,
@@ -431,6 +432,7 @@ test("session methods use the public HTTP contract", async () => {
     sessionID: "ses_test",
     model: { id: "claude", providerID: "anthropic" },
   })
+  await client.session.archive({ sessionID: "ses_test" })
   const admitted = await client.session.prompt({
     sessionID: "ses_test",
     text: "Hello",
@@ -466,6 +468,7 @@ test("session methods use the public HTTP contract", async () => {
     ["POST", "http://localhost:3000/api/session"],
     ["POST", "http://localhost:3000/api/session/ses_test/agent"],
     ["POST", "http://localhost:3000/api/session/ses_test/model"],
+    ["POST", "http://localhost:3000/api/session/ses_test/archive"],
     ["POST", "http://localhost:3000/api/session/ses_test/prompt"],
     ["POST", "http://localhost:3000/api/session/ses_test/generate"],
     ["POST", "http://localhost:3000/api/session/ses_test/synthetic"],
@@ -546,6 +549,7 @@ const session = {
 
 const admission = {
   data: {
+    admittedSeq: 0,
     id: "msg_test",
     sessionID: "ses_test",
     type: "user",
@@ -557,6 +561,7 @@ const admission = {
 
 const syntheticAdmission = {
   data: {
+    admittedSeq: 1,
     id: "msg_synthetic",
     sessionID: "ses_test",
     type: "synthetic",
@@ -569,6 +574,7 @@ const syntheticAdmission = {
 const compactionAdmission = {
   data: {
     type: "compaction",
+    admittedSeq: 1,
     id: "msg_compaction",
     sessionID: "ses_test",
     timeCreated: 1_717_171_717_000,

packages/core/package.json

@@ -92,7 +92,6 @@
     "@lydell/node-pty": "catalog:",
     "@modelcontextprotocol/sdk": "1.29.0",
     "@ff-labs/fff-bun": "0.10.1",
-    "@ff-labs/fff-node": "0.10.1",
     "@opencode-ai/codemode": "workspace:*",
     "@opencode-ai/effect-drizzle-sqlite": "workspace:*",
     "@opencode-ai/effect-sqlite-node": "workspace:*",

packages/core/schema.json

@@ -1,9 +1,9 @@
 {
   "version": "7",
   "dialect": "sqlite",
-  "id": "db37a97f-9b5e-4c87-be8b-4feace35136c",
+  "id": "a4ba73b4-21bc-41ab-a415-94e2ca38d798",
   "prevIds": [
-    "a4ba73b4-21bc-41ab-a415-94e2ca38d798"
+    "5f0a1db8-d4bf-42c3-becb-96b46fe66bed"
   ],
   "ddl": [
     {
@@ -1266,7 +1266,17 @@
       "autoincrement": false,
       "default": null,
       "generated": null,
-      "name": "fork_boundary",
+      "name": "fork_message_id",
+      "entityType": "columns",
+      "table": "session"
+    },
+    {
+      "type": "integer",
+      "notNull": false,
+      "autoincrement": false,
+      "default": null,
+      "generated": null,
+      "name": "fork_seq",
       "entityType": "columns",
       "table": "session"
     },

packages/core/src/codemode/instructions.ts

@@ -6,9 +6,7 @@ import { Instructions } from "../instructions/index"
 import { CodeModeCatalog } from "./catalog"
 
 // prettier-ignore
-const prompt = (hasMoreTools: boolean) => `The Code Mode tool catalog below is ${hasMoreTools ? "partial" : "complete"}.
-
-${hasMoreTools ? "The Code Mode catalog and `search` results are" : "This catalog is"} the complete set of tools available within Code Mode. Tools presented elsewhere are not available in this runtime.${hasMoreTools ? `
+const prompt = (hasMoreTools: boolean) => `The Code Mode tool catalog below is ${hasMoreTools ? "partial" : "complete"}.${hasMoreTools ? `
 
 ## Search
 

packages/core/src/codemode/tool.ts

@@ -33,7 +33,7 @@ type CollectedFiles = {
 
 // Invariant model-facing guidance; the changing tool catalog is delivered through Instructions.
 const description = [
-  "Run JavaScript in a confined Code Mode runtime to orchestrate tool calls and compose their results.",
+  "Run JavaScript to orchestrate tool calls and compose their results through `{ code }` in a confined Code Mode runtime.",
   "Imports, direct filesystem access, and timers are unavailable. Do not use `fetch`; all external access goes through `tools`.",
   "Within `{ code }`, the only callable tools are those explicitly listed in the Code Mode catalog instructions or returned by `search`. Inside `{ code }`, ignore tools shown outside the Code Mode catalog. They are not available in the Code Mode runtime.",
   'Call tools through `tools` using only exact paths and signatures from the catalog. Do not infer or normalize tool names; preserve bracket notation such as `tools.<namespace>["tool-name"](input)`.',

packages/core/src/config.ts

@@ -393,8 +393,7 @@ export const layer = (options?: Options) => Layer.effect(
         const key = JSON.stringify(target)
         if (watched.has(key)) continue
         watched.add(key)
-        const stream = yield* watcher.subscribe(target)
-        yield* stream.pipe(
+        yield* watcher.subscribe(target).pipe(
           Stream.runForEach((update) => PubSub.publish(updates, update)),
           Effect.forkScoped({ startImmediately: true }),
         )

packages/core/src/database/migration.gen.ts

@@ -57,6 +57,5 @@ export const migrations = (
     import("./migration/20260716020354_kv"),
     import("./migration/20260722011141_delete_tool_progress_events"),
     import("./migration/20260722170000_canonical_tool_results"),
-    import("./migration/20260729022634_session_fork_boundary"),
   ])
 ).map((module) => module.default) satisfies DatabaseMigration.Migration[]

packages/core/src/database/migration/20260729022634_session_fork_boundary.ts

@@ -1,13 +0,0 @@
-import { Effect } from "effect"
-import type { DatabaseMigration } from "../migration"
-
-export default {
-  id: "20260729022634_session_fork_boundary",
-  up(tx) {
-    return Effect.gen(function* () {
-      yield* tx.run(`ALTER TABLE \`session\` ADD \`fork_boundary\` text;`)
-      yield* tx.run(`ALTER TABLE \`session\` DROP COLUMN \`fork_message_id\`;`)
-      yield* tx.run(`ALTER TABLE \`session\` DROP COLUMN \`fork_seq\`;`)
-    })
-  },
-} satisfies DatabaseMigration.Migration

packages/core/src/database/schema.gen.ts

@@ -213,7 +213,8 @@ export default {
           \`workspace_id\` text,
           \`parent_id\` text,
           \`fork_session_id\` text,
-          \`fork_boundary\` text,
+          \`fork_message_id\` text,
+          \`fork_seq\` integer,
           \`slug\` text NOT NULL,
           \`directory\` text NOT NULL,
           \`path\` text,

packages/core/src/filesystem/fff.node.ts

@@ -1,52 +1,71 @@
-import type {
-  DirItem,
-  DirSearchResult,
-  FileItem,
-  GrepCursor,
-  GrepMatch,
-  GrepResult,
-  InitOptions,
-  MixedItem,
-  MixedSearchResult,
-  SearchResult,
-} from "@ff-labs/fff-node"
-
-const { FileFinder } = await import("@ff-labs/fff-node").catch(() => ({ FileFinder: undefined }))
-
 export type Result<T> = { ok: true; value: T } | { ok: false; error: string }
 
-export type Init = InitOptions
+export interface Init {
+  basePath: string
+  frecencyDbPath?: string
+  historyDbPath?: string
+  useUnsafeNoLock?: boolean
+  disableMmapCache?: boolean
+  disableContentIndexing?: boolean
+  disableWatch?: boolean
+  aiMode?: boolean
+  logFilePath?: string
+  logLevel?: "trace" | "debug" | "info" | "warn" | "error"
+  enableFsRootScanning?: boolean
+  enableHomeDirScanning?: boolean
+}
+
+export interface File {
+  relativePath: string
+  fileName: string
+  modified: number
+}
+
+export interface Directory {
+  relativePath: string
+  dirName: string
+  maxAccessFrecency: number
+}
+
+export type Mixed = { type: "file"; item: File } | { type: "directory"; item: Directory }
 
 export interface Search {
-  items: FileItem[]
-  scores: SearchResult["scores"]
+  items: File[]
+  scores: Array<{ total: number }>
   totalMatched: number
   totalFiles: number
 }
 
 export interface DirSearch {
-  items: DirItem[]
-  scores: DirSearchResult["scores"]
+  items: Directory[]
+  scores: Array<{ total: number }>
   totalMatched: number
   totalDirs: number
 }
 
 export interface MixedSearch {
-  items: MixedItem[]
-  scores: MixedSearchResult["scores"]
+  items: Mixed[]
+  scores: Array<{ total: number }>
   totalMatched: number
   totalFiles: number
   totalDirs: number
 }
 
-export type File = FileItem
-export type Directory = DirItem
-export type Mixed = MixedItem
-export type Cursor = GrepCursor | null
-export type Hit = GrepMatch
+export type Cursor = null
+
+export interface Hit {
+  relativePath: string
+  fileName: string
+  lineNumber: number
+  byteOffset: number
+  lineContent: string
+  matchRanges: [number, number][]
+  contextBefore?: string[]
+  contextAfter?: string[]
+}
 
 export interface Grep {
-  items: GrepResult["items"]
+  items: Hit[]
   totalMatched: number
   totalFilesSearched: number
   totalFiles: number
@@ -109,30 +128,11 @@ export interface Picker {
 }
 
 export function available() {
-  return FileFinder?.isAvailable() ?? false
+  return false
 }
 
-export function create(opts: Init): Result<Picker> {
-  if (!FileFinder) return { ok: false, error: "fff unavailable on node runtime" }
-  const made = FileFinder.create(opts)
-  if (!made.ok) return made
-  const pick = made.value
-  return {
-    ok: true,
-    value: {
-      destroy: () => pick.destroy(),
-      isScanning: () => pick.isScanning(),
-      waitForScan: (timeoutMs) => pick.waitForScan(timeoutMs),
-      refreshGitStatus: () => pick.refreshGitStatus(),
-      fileSearch: (query, next) => pick.fileSearch(query, next),
-      glob: (pattern, next) => pick.glob(pattern, next),
-      directorySearch: (query, next) => pick.directorySearch(query, next),
-      mixedSearch: (query, next) => pick.mixedSearch(query, next),
-      grep: (query, next) => pick.grep(query, next),
-      trackQuery: (query, file) => pick.trackQuery(query, file),
-      getHistoricalQuery: (offset) => pick.getHistoricalQuery(offset),
-    },
-  }
+export function create(_opts: Init): Result<Picker> {
+  return { ok: false, error: "fff unavailable on node runtime" }
 }
 
 export * as Fff from "./fff.node"