Release 0.2.2 (#638 )

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
refactor: migrate to Nextjs Route Handler (#625 )
2026-07-02 19:14:28 -04:00 · 2025-05-22 17:17:34 +07:00 · 2025-05-22 11:47:24 +07:00 · 2025-05-21 14:31:41 +07:00 · 2025-05-21 12:49:44 +07:00 · 2025-05-21 12:31:11 +07:00
82 changed files with 4580 additions and 2997 deletions
@@ -1,5 +0,0 @@
---
-"create-llama": patch
---
-
-Fix typing check issue
@@ -1,5 +0,0 @@
---
-"create-llama": patch
---
-
-fix chromadb dependency issue
@@ -1,5 +0,0 @@
---
-"create-llama": patch
---
-
-fix: remove dead generated ai code
@@ -1,5 +0,0 @@
---
-"create-llama": patch
---
-
-Deprecate pro mode
@@ -79,7 +79,7 @@ jobs:
      - uses: actions/upload-artifact@v4
        if: always()
        with:
-          name: playwright-report-python-${{ matrix.os }}-${{ matrix.frameworks }}-${{ matrix.datasources }}
+          name: playwright-report-python-${{ matrix.os }}-${{ matrix.frameworks }}-${{ matrix.datasources }}-${{ matrix.template-types }}
          path: packages/create-llama/playwright-report/
          overwrite: true
          retention-days: 30
@@ -136,6 +136,21 @@ jobs:
        run: pnpm run pack-install
        working-directory: packages/create-llama

+      - name: Build server
+        run: pnpm run build
+        working-directory: packages/server
+
+      - name: Pack @llamaindex/server package
+        run: |
+          pnpm pack --pack-destination "${{ runner.temp }}"
+          if [ "${{ runner.os }}" == "Windows" ]; then
+            file=$(find "${{ runner.temp }}" -name "llamaindex-server-*.tgz" | head -n 1)
+            mv "$file" "${{ runner.temp }}/llamaindex-server.tgz"
+          else
+            mv ${{ runner.temp }}/llamaindex-server-*.tgz ${{ runner.temp }}/llamaindex-server.tgz
+          fi
+        working-directory: packages/server
+
      - name: Run Playwright tests for TypeScript
        run: pnpm run e2e:typescript
        env:
@@ -144,12 +159,13 @@ jobs:
          FRAMEWORK: ${{ matrix.frameworks }}
          DATASOURCE: ${{ matrix.datasources }}
          TEMPLATE_TYPE: ${{ matrix.template-types }}
+          SERVER_PACKAGE_PATH: ${{ runner.temp }}/llamaindex-server.tgz
        working-directory: packages/create-llama

      - uses: actions/upload-artifact@v4
        if: always()
        with:
-          name: playwright-report-typescript-${{ matrix.os }}-${{ matrix.frameworks }}-${{ matrix.datasources }}-node${{ matrix.node-version }}
+          name: playwright-report-typescript-${{ matrix.os }}-${{ matrix.frameworks }}-${{ matrix.datasources }}-node${{ matrix.node-version }}-${{ matrix.template-types }}
          path: packages/create-llama/playwright-report/
          overwrite: true
          retention-days: 30
@@ -1,5 +1,29 @@
 # create-llama

+## 0.5.16
+
+### Patch Changes
+
+- 6f75d4a: fix: unsupported language in code gen workflow
+- d0618fa: Fix LlamaCloud generate script issue
+
+## 0.5.15
+
+### Patch Changes
+
+- 527075c: Enable dev mode that allows updating code directly in the UI
+
+## 0.5.14
+
+### Patch Changes
+
+- 1df8cfb: Split artifacts use case to document generator and code generator
+- 1b5a519: chore: improve dev experience with nodemon
+- b3eb0ba: Fix typing check issue
+- 556f33c: fix chromadb dependency issue
+- 2451539: fix: remove dead generated ai code
+- 7a70390: Deprecate pro mode
+
 ## 0.5.13

 ### Patch Changes
@@ -18,7 +18,8 @@ const useCases: TemplateUseCase[] = [
  "agentic_rag",
  "deep_research",
  "financial_report",
-  "artifacts",
+  "code_generator",
+  "document_generator",
 ];
 const dataSource: string = process.env.DATASOURCE
  ? process.env.DATASOURCE
@@ -12,21 +12,30 @@ import { createTestDir, runCreateLlama, type AppType } from "../utils";
 const templateFramework: TemplateFramework = process.env.FRAMEWORK
  ? (process.env.FRAMEWORK as TemplateFramework)
  : "fastapi";
-const dataSource: string = "--example-file";
+const dataSource: string = process.env.DATASOURCE
+  ? (process.env.DATASOURCE as string)
+  : "--example-file";
+const llamaCloudProjectName = "create-llama";
+const llamaCloudIndexName = "e2e-test";
+
 const templateUI: TemplateUI = "shadcn";
 const templatePostInstallAction: TemplatePostInstallAction = "runApp";
 const appType: AppType = "--frontend";
 const userMessage = "Write a blog post about physical standards for letters";
-const templateUseCases = ["financial_report", "agentic_rag", "deep_research"];
+const templateUseCases = [
+  "agentic_rag",
+  "financial_report",
+  "deep_research",
+  "code_generator",
+];

 for (const useCase of templateUseCases) {
  test.describe(`Test use case ${useCase} ${templateFramework} ${dataSource} ${templateUI} ${appType} ${templatePostInstallAction}`, async () => {
    test.skip(
-      process.platform !== "linux" ||
-        process.env.DATASOURCE === "--no-files" ||
-        templateFramework === "express",
+      dataSource === "--no-files" || templateFramework === "express",
      "The llamaindexserver template currently only works with nextjs, fastapi. We also only run on Linux to speed up tests.",
    );
+    const useLlamaParse = dataSource === "--llamacloud";
    let port: number;
    let cwd: string;
    let name: string;
@@ -48,6 +57,9 @@ for (const useCase of templateUseCases) {
        templateUI,
        appType,
        useCase,
+        llamaCloudProjectName,
+        llamaCloudIndexName,
+        useLlamaParse,
      });
      name = result.projectName;
      appProcess = result.appProcess;
@@ -23,7 +23,8 @@ const useCases: TemplateUseCase[] = [
  "agentic_rag",
  "deep_research",
  "financial_report",
-  "artifacts",
+  "code_generator",
+  "document_generator",
 ];
 const dataSource: string = process.env.DATASOURCE
  ? process.env.DATASOURCE
@@ -82,7 +83,7 @@ test.describe("Test resolve TS dependencies", () => {
          });
        });
        // Skipping llamacloud for the use case doesn't use index.
-        if (useCase !== "artifacts") {
+        if (useCase !== "code_generator" && useCase !== "document_generator") {
          test(`llamaParse - ${optionDescription}`, async () => {
            await runTest({
              templateType: templateType,
@@ -18,6 +18,7 @@ import {
  ModelConfig,
  TemplateDataSource,
  TemplateFramework,
+  TemplateUseCase,
  TemplateVectorDB,
 } from "./types";
 import { installTSTemplate } from "./typescript";
@@ -60,6 +61,7 @@ async function generateContextData(
  vectorDb?: TemplateVectorDB,
  llamaCloudKey?: string,
  useLlamaParse?: boolean,
+  useCase?: TemplateUseCase,
 ) {
  if (packageManager) {
    const runGenerate = `${cyan(
@@ -96,7 +98,12 @@ async function generateContextData(
        }
      } else {
        console.log(`Running ${runGenerate} to generate the context data.`);
-        await callPackageManager(packageManager, true, ["run", "generate"]);
+        const shouldRunGenerate =
+          useCase !== "code_generator" && useCase !== "document_generator"; // Artifact use case doesn't use index.
+
+        if (shouldRunGenerate) {
+          await callPackageManager(packageManager, true, ["run", "generate"]);
+        }
        return;
      }
    }
@@ -224,6 +231,7 @@ export const installTemplate = async (
        props.vectorDb,
        props.llamaCloudKey,
        props.useLlamaParse,
+        props.useCase,
      );
    }

@@ -267,7 +267,7 @@ const getAdditionalDependencies = (
    if (observability === "traceloop") {
      dependencies.push({
        name: "traceloop-sdk",
-        version: ">=0.15.11,<0.16.0",
+        version: ">=0.15.11",
      });
    }
    if (observability === "llamatrace") {
@@ -677,6 +677,7 @@ export const installPythonTemplate = async ({
    dataSources,
    tools,
    template,
+    observability,
  );

  await addDependencies(root, addOnDependencies);
@@ -58,7 +58,8 @@ export type TemplateUseCase =
  | "extractor"
  | "contract_review"
  | "agentic_rag"
-  | "artifacts";
+  | "code_generator"
+  | "document_generator";
 // Config for both file and folder
 export type FileSourceConfig =
  | {
@@ -77,7 +77,7 @@ const installLlamaIndexServerTemplate = async ({
  }

  // Simplify use case code
-  if (useCase === "artifacts") {
+  if (useCase === "code_generator" || useCase === "document_generator") {
    // Artifact use case doesn't use index.
    // We don't need data.ts, generate.ts
    await fs.rm(path.join(root, "src", "app", "data.ts"));
@@ -387,7 +387,7 @@ const providerDependencies: {
  [key in ModelProvider]?: Record<string, string>;
 } = {
  openai: {
-    "@llamaindex/openai": "^0.2.0",
+    "@llamaindex/openai": "~0.4.0",
  },
  gemini: {
    "@llamaindex/google": "^0.2.0",
@@ -513,7 +513,7 @@ async function updatePackageJson({
  if (backend) {
    packageJson.dependencies = {
      ...packageJson.dependencies,
-      "@llamaindex/readers": "^3.0.0",
+      "@llamaindex/readers": "~3.1.4",
    };

    if (vectorDb && vectorDb in vectorDbDependencies) {
@@ -543,6 +543,16 @@ async function updatePackageJson({
    };
  }

+  // if having custom server package tgz file, use it for testing @llamaindex/server
+  const serverPackagePath = process.env.SERVER_PACKAGE_PATH;
+  if (serverPackagePath && template === "llamaindexserver") {
+    const relativePath = path.relative(process.cwd(), serverPackagePath);
+    packageJson.dependencies = {
+      ...packageJson.dependencies,
+      "@llamaindex/server": `file:${relativePath}`,
+    };
+  }
+
  await fs.writeFile(
    packageJsonFile,
    JSON.stringify(packageJson, null, 2) + os.EOL,
@@ -1,6 +1,6 @@
 {
  "name": "create-llama",
-  "version": "0.5.13",
+  "version": "0.5.16",
  "description": "Create LlamaIndex-powered apps with one command",
  "keywords": [
    "rag",
@@ -6,7 +6,7 @@ const defaults: Omit<QuestionArgs, "modelConfig"> = {
  framework: "nextjs",
  ui: "shadcn",
  frontend: false,
-  llamaCloudKey: "",
+  llamaCloudKey: undefined,
  useLlamaParse: false,
  communityProjectConfig: undefined,
  llamapack: "",
@@ -10,7 +10,8 @@ type AppType =
  | "agentic_rag"
  | "financial_report"
  | "deep_research"
-  | "artifacts";
+  | "code_generator"
+  | "document_generator";

 type SimpleAnswers = {
  appType: AppType;
@@ -47,10 +48,14 @@ export const askSimpleQuestions = async (
            "Researches and analyzes provided documents from multiple perspectives, generating a comprehensive report with citations to support key findings and insights.",
        },
        {
-          title: "Artifacts",
-          value: "artifacts",
-          description:
-            "Build your own Vercel's v0 or OpenAI's canvas-styled UI.",
+          title: "Code Generator",
+          value: "code_generator",
+          description: "Build a Vercel v0 styled code generator.",
+        },
+        {
+          title: "Document Generator",
+          value: "document_generator",
+          description: "Build a OpenAI canvas-styled document generator.",
        },
      ],
    },
@@ -76,7 +81,7 @@ export const askSimpleQuestions = async (
  );
  language = newLanguage;

-  if (appType !== "artifacts") {
+  if (appType !== "code_generator" && appType !== "document_generator") {
    const { useLlamaCloud: newUseLlamaCloud } = await prompts(
      {
        type: "toggle",
@@ -153,7 +158,13 @@ const convertAnswers = async (
      tools: [],
      modelConfig: MODEL_GPT41,
    },
-    artifacts: {
+    code_generator: {
+      template: "llamaindexserver",
+      dataSources: [],
+      tools: [],
+      modelConfig: MODEL_GPT41,
+    },
+    document_generator: {
      template: "llamaindexserver",
      dataSources: [],
      tools: [],
@@ -12,7 +12,7 @@ dependencies = [
    "llama-index>=0.12.1",
    "llama-parse>=0.6.21,<0.7.0",
    "cachetools>=5.3.3",
-    "reflex>=0.6.2.post1",
+    "reflex==0.7.10",
 ]

 [project.scripts]
@@ -13,7 +13,7 @@ dependencies = [
    "llama-index>=0.12.1",
    "llama-parse>=0.6.21,<0.7.0",
    "cachetools>=5.3.3",
-    "reflex>=0.6.2.post1",
+    "reflex==0.7.10",
 ]

 [project.scripts]
@@ -0,0 +1,132 @@
+import { Badge } from "@/components/ui/badge";
+import { Card, CardContent, CardHeader, CardTitle } from "@/components/ui/card";
+import { Progress } from "@/components/ui/progress";
+import { Skeleton } from "@/components/ui/skeleton";
+import { cn } from "@/lib/utils";
+import { Markdown } from "@llamaindex/chat-ui/widgets";
+import { ListChecks, Loader2, Wand2 } from "lucide-react";
+import { useEffect, useState } from "react";
+
+const STAGE_META = {
+  plan: {
+    icon: ListChecks,
+    badgeText: "Step 1/2: Planning",
+    gradient: "from-blue-100 via-blue-50 to-white",
+    progress: 33,
+    iconBg: "bg-blue-100 text-blue-600",
+    badge: "bg-blue-100 text-blue-700",
+  },
+  generate: {
+    icon: Wand2,
+    badgeText: "Step 2/2: Generating",
+    gradient: "from-violet-100 via-violet-50 to-white",
+    progress: 66,
+    iconBg: "bg-violet-100 text-violet-600",
+    badge: "bg-violet-100 text-violet-700",
+  },
+};
+
+function ArtifactWorkflowCard({ event }) {
+  const [visible, setVisible] = useState(event?.state !== "completed");
+  const [fade, setFade] = useState(false);
+
+  useEffect(() => {
+    if (event?.state === "completed") {
+      setVisible(false);
+    } else {
+      setVisible(true);
+      setFade(false);
+    }
+  }, [event?.state]);
+
+  if (!event || !visible) return null;
+
+  const { state, requirement } = event;
+  const meta = STAGE_META[state];
+
+  if (!meta) return null;
+
+  return (
+    <div className="flex min-h-[180px] w-full items-center justify-center py-2">
+      <Card
+        className={cn(
+          "w-full rounded-xl shadow-md transition-all duration-500",
+          "border-0",
+          fade && "pointer-events-none opacity-0",
+          `bg-gradient-to-br ${meta.gradient}`,
+        )}
+        style={{
+          boxShadow:
+            "0 2px 12px 0 rgba(80, 80, 120, 0.08), 0 1px 3px 0 rgba(80, 80, 120, 0.04)",
+        }}
+      >
+        <CardHeader className="flex flex-row items-center gap-2 px-3 pb-1 pt-2">
+          <div
+            className={cn(
+              "flex items-center justify-center rounded-full p-1",
+              meta.iconBg,
+            )}
+          >
+            <meta.icon className="h-5 w-5" />
+          </div>
+          <CardTitle className="flex items-center gap-2 text-base font-semibold">
+            <Badge className={cn("ml-1", meta.badge, "px-2 py-0.5 text-xs")}>
+              {meta.badgeText}
+            </Badge>
+          </CardTitle>
+        </CardHeader>
+        <CardContent className="px-3 py-1">
+          {state === "plan" && (
+            <div className="flex flex-col items-center gap-2 py-2">
+              <Loader2 className="mb-1 h-6 w-6 animate-spin text-blue-400" />
+              <div className="text-center text-sm font-medium text-blue-900">
+                Analyzing your request...
+              </div>
+              <Skeleton className="mt-1 h-3 w-1/2 rounded-full" />
+            </div>
+          )}
+          {state === "generate" && (
+            <div className="flex flex-col gap-2 py-2">
+              <div className="flex items-center gap-1">
+                <Loader2 className="h-4 w-4 animate-spin text-violet-400" />
+                <span className="text-sm font-medium text-violet-900">
+                  Working on the requirement:
+                </span>
+              </div>
+              <div className="max-h-24 overflow-auto rounded-lg border border-violet-200 bg-violet-50 px-2 py-1 text-xs">
+                {requirement ? (
+                  <Markdown content={requirement} />
+                ) : (
+                  <span className="italic text-violet-400">
+                    No requirements available yet.
+                  </span>
+                )}
+              </div>
+            </div>
+          )}
+        </CardContent>
+        <div className="px-3 pb-2 pt-1">
+          <Progress
+            value={meta.progress}
+            className={cn(
+              "h-1 rounded-full bg-gray-200",
+              state === "plan" && "bg-blue-200",
+              state === "generate" && "bg-violet-200",
+            )}
+          />
+        </div>
+      </Card>
+    </div>
+  );
+}
+
+export default function Component({ events }) {
+  const aggregateEvents = () => {
+    if (!events || events.length === 0) return null;
+    return events[events.length - 1];
+  };
+
+  const event = aggregateEvents();
+
+  return <ArtifactWorkflowCard event={event} />;
+}
@@ -1,15 +0,0 @@
-from app.code_workflow import CodeArtifactWorkflow
-
-# from app.document_workflow import DocumentArtifactWorkflow to generate documents
-from llama_index.core.workflow import Workflow
-from llama_index.llms.openai import OpenAI
-from llama_index.server.api.models import ChatRequest
-
-
-def create_workflow(chat_request: ChatRequest) -> Workflow:
-    workflow = CodeArtifactWorkflow(
-        llm=OpenAI(model="gpt-4.1"),
-        chat_request=chat_request,
-        timeout=120.0,
-    )
-    return workflow
@@ -0,0 +1,65 @@
+This is a [LlamaIndex](https://www.llamaindex.ai/) project using [Workflows](https://docs.llamaindex.ai/en/stable/understanding/workflows/).
+
+## Getting Started
+
+First, setup the environment with uv:
+
+> **_Note:_** This step is not needed if you are using the dev-container.
+
+```shell
+uv sync
+```
+
+Then check the parameters that have been pre-configured in the `.env` file in this directory.
+Make sure you have set the `OPENAI_API_KEY` for the LLM.
+
+Then, run the development server:
+
+```shell
+uv run fastapi dev
+```
+
+Then open [http://localhost:8000](http://localhost:8000) with your browser to start the chat UI.
+
+To start the app optimized for **production**, run:
+
+```
+uv run fastapi run
+```
+
+## Configure LLM and Embedding Model
+
+You can configure [LLM model](https://docs.llamaindex.ai/en/stable/module_guides/models/llms) and [embedding model](https://docs.llamaindex.ai/en/stable/module_guides/models/embeddings) in [settings.py](app/settings.py).
+
+## Use Case
+AI-powered code generator that can help you generate app with a chat interface, code editor and app preview.
+
+To update the workflow, you can modify the code in [`workflow.py`](app/workflow.py).
+
+You can start by sending an request on the [chat UI](http://localhost:8000) or you can test the `/api/chat` endpoint with the following curl request:
+
+```
+curl --location 'localhost:8000/api/chat' \
+--header 'Content-Type: application/json' \
+--data '{ "messages": [{ "role": "user", "content": "Create a report comparing the finances of Apple and Tesla" }] }'
+```
+
+## Customize the UI
+
+To customize the UI, you can start by modifying the [./components/ui_event.jsx](./components/ui_event.jsx) file.
+
+You can also generate a new code for the workflow using LLM by running the following command:
+
+```
+uv run generate_ui
+```
+
+## Learn More
+
+To learn more about LlamaIndex, take a look at the following resources:
+
+- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex.
+- [Workflows Introduction](https://docs.llamaindex.ai/en/stable/understanding/workflows/) - learn about LlamaIndex workflows.
+- [LlamaIndex Server](https://pypi.org/project/llama-index-server/)
+
+You can check out [the LlamaIndex GitHub repository](https://github.com/run-llama/llama_index) - your feedback and contributions are welcome!
@@ -6,6 +6,7 @@ from llama_index.core.chat_engine.types import ChatMessage
 from llama_index.core.llms import LLM
 from llama_index.core.memory import ChatMemoryBuffer
 from llama_index.core.prompts import PromptTemplate
+from llama_index.llms.openai import OpenAI
 from llama_index.core.workflow import (
    Context,
    Event,
@@ -26,6 +27,15 @@ from llama_index.server.api.utils import get_last_artifact
 from pydantic import BaseModel, Field


+def create_workflow(chat_request: ChatRequest) -> Workflow:
+    workflow = CodeArtifactWorkflow(
+        llm=OpenAI(model="gpt-4.1"),
+        chat_request=chat_request,
+        timeout=120.0,
+    )
+    return workflow
+
+
 class Requirement(BaseModel):
    next_step: Literal["answering", "coding"]
    language: Optional[str] = None
@@ -33,12 +33,9 @@ You can configure [LLM model](https://docs.llamaindex.ai/en/stable/module_guides

 ## Use Case

-We have prepared two artifact workflows:
+AI-powered document generator that can help you generate documents with a chat interface and simple markdown editor.

- [Code Workflow](app/code_workflow.py): To generate code and display it in the UI like Vercel's v0.
- [Document Workflow](app/document_workflow.py): Generate and update a document like OpenAI's canvas.
-
-Modify the factory method in [`workflow.py`](app/workflow.py) to decide which artifact workflow to use. Without any changes the Code Workflow is used.
+To update the workflow, you can modify the code in [`workflow.py`](app/workflow.py).

 You can start by sending an request on the [chat UI](http://localhost:8000) or you can test the `/api/chat` endpoint with the following curl request:

@@ -4,6 +4,7 @@ from typing import Any, Literal, Optional

 from llama_index.core.chat_engine.types import ChatMessage
 from llama_index.core.llms import LLM
+from llama_index.llms.openai import OpenAI
 from llama_index.core.memory import ChatMemoryBuffer
 from llama_index.core.prompts import PromptTemplate
 from llama_index.core.workflow import (
@@ -26,6 +27,15 @@ from llama_index.server.api.utils import get_last_artifact
 from pydantic import BaseModel, Field


+def create_workflow(chat_request: ChatRequest) -> Workflow:
+    workflow = DocumentArtifactWorkflow(
+        llm=OpenAI(model="gpt-4.1"),
+        chat_request=chat_request,
+        timeout=120.0,
+    )
+    return workflow
+
+
 class DocumentRequirement(BaseModel):
    type: Literal["markdown", "html"]
    title: str
@@ -1,12 +0,0 @@
-import { createCodeArtifactWorkflow, UIEventSchema } from "./code-workflow";
-// import { createDocumentArtifactWorkflow, UIEventSchema } from "./doc-workflow";
-
-export const workflowFactory = async (reqBody: any) => {
-  // Uncomment the import and change to createDocumentArtifactWorkflow to use the document workflow
-  const workflow = createCodeArtifactWorkflow(reqBody);
-
-  return workflow;
-};
-
-// Re-export the UIEventSchema for generating the UI by `pnpm generate:ui` command
-export { UIEventSchema };
@@ -30,12 +30,9 @@ npm run generate:ui

 ## Use Case

-We have prepared two artifact workflows:
+AI-powered code generator that can help you generate app with a chat interface, code editor and app preview.

- [Code Workflow](app/code_workflow.ts): To generate code and display it in the UI like Vercel's v0.
- [Document Workflow](app/document_workflow.ts): Generate and update a document like OpenAI's canvas.
-
-Modify the factory method in [`workflow.ts`](app/workflow.ts) to decide which artifact workflow to use. Without any changes the Code Workflow is used.
+To update the workflow, you can modify the code in [`workflow.ts`](app/workflow.ts).

 You can start by sending an request on the [chat UI](http://localhost:3000) or you can test the `/api/chat` endpoint with the following curl request:

@@ -1,5 +1,5 @@
 import { extractLastArtifact } from "@llamaindex/server";
-import { ChatMemoryBuffer, LLM, Settings } from "llamaindex";
+import { ChatMemoryBuffer, MessageContent, Settings } from "llamaindex";

 import {
  agentStreamEvent,
@@ -40,7 +40,7 @@ export const UIEventSchema = z.object({

 export type UIEvent = z.infer<typeof UIEventSchema>;
 const planEvent = workflowEvent<{
-  userInput: string;
+  userInput: MessageContent;
  context?: string | undefined;
 }>();

@@ -65,34 +65,33 @@ const artifactEvent = workflowEvent<{
  };
 }>();

-export function createCodeArtifactWorkflow(reqBody: any, llm?: LLM) {
-  if (!llm) {
-    llm = Settings.llm;
-  }
+export function workflowFactory(reqBody: any) {
+  const llm = Settings.llm;
+
  const { withState, getContext } = createStatefulMiddleware(() => {
    return {
-      memory: new ChatMemoryBuffer({
-        llm,
-        chatHistory: reqBody.chatHistory,
-      }),
+      memory: new ChatMemoryBuffer({ llm }),
      lastArtifact: extractLastArtifact(reqBody),
    };
  });
  const workflow = withState(createWorkflow());

-  workflow.handle([startAgentEvent], async ({ data: { userInput } }) => {
+  workflow.handle([startAgentEvent], async ({ data }) => {
+    const { userInput, chatHistory = [] } = data;
    // Prepare chat history
    const { state } = getContext();
    // Put user input to the memory
    if (!userInput) {
      throw new Error("Missing user input to start the workflow");
    }
-    state.memory.put({
-      role: "user",
-      content: userInput,
-    });
+    state.memory.set(chatHistory);
+    state.memory.put({ role: "user", content: userInput });
+
    return planEvent.with({
-      userInput,
+      userInput: userInput,
+      context: state.lastArtifact
+        ? JSON.stringify(state.lastArtifact)
+        : undefined,
    });
  });

@@ -177,15 +176,6 @@ ${user_msg}
      throw new Error("No JSON block found in the response.");
    }
    const requirement = RequirementSchema.parse(JSON.parse(jsonBlock[1]));
-    sendEvent(
-      uiEvent.with({
-        type: "ui_event",
-        data: {
-          state: "generate",
-          requirement: requirement.requirement,
-        },
-      }),
-    );
    state.memory.put({
      role: "assistant",
      content: `The plan for next step: \n${response.text}`,
@@ -10,12 +10,14 @@ import {
 import {
  ChatMemoryBuffer,
  LlamaCloudIndex,
+  MessageContent,
  Metadata,
  MetadataMode,
  NodeWithScore,
  PromptTemplate,
  Settings,
  VectorStoreIndex,
+  extractText,
 } from "llamaindex";
 import { randomUUID } from "node:crypto";
 import { z } from "zod";
@@ -153,7 +155,7 @@ export function getWorkflow(index: VectorStoreIndex | LlamaCloudIndex) {
        chatHistory: [],
      }),
      contextNodes: [] as NodeWithScore<Metadata>[],
-      userRequest: "",
+      userRequest: "" as MessageContent,
      totalQuestions: 0,
      researchResults: [] as ResearchResult[],
    };
@@ -178,7 +180,7 @@ export function getWorkflow(index: VectorStoreIndex | LlamaCloudIndex) {
      }),
    );

-    const retrievedNodes = await retriever.retrieve(userInput);
+    const retrievedNodes = await retriever.retrieve({ query: userInput });

    sendEvent(toSourceEvent(retrievedNodes));
    sendEvent(
@@ -349,7 +351,7 @@ const createResearchPlan = async (
  memory: ChatMemoryBuffer,
  contextStr: string,
  enhancedPrompt: string,
-  userRequest: string,
+  userRequest: MessageContent,
 ) => {
  const chatHistory = await memory.getMessages();

@@ -361,7 +363,7 @@ const createResearchPlan = async (
    context_str: contextStr,
    conversation_context: conversationContext,
    enhanced_prompt: enhancedPrompt,
-    user_request: userRequest,
+    user_request: extractText(userRequest),
  });

  const responseFormat = z.object({
@@ -0,0 +1,53 @@
+This is a [LlamaIndex](https://www.llamaindex.ai/) project bootstrapped with [`create-llama`](https://github.com/run-llama/LlamaIndexTS/tree/main/packages/create-llama).
+
+## Getting Started
+
+First, install the dependencies:
+
+```
+npm install
+```
+
+Third, run the development server:
+
+```
+npm run dev
+```
+
+Open [http://localhost:3000](http://localhost:3000) with your browser to see the chat UI.
+
+## Configure LLM and Embedding Model
+
+You can configure [LLM model](https://ts.llamaindex.ai/docs/llamaindex/modules/llms) in the [settings file](src/app/settings.ts).
+
+## Custom UI Components
+
+We have a custom component located in `components/ui_event.jsx`. This is used to display the state of artifact workflows in UI. You can regenerate a new UI component from the workflow event schema by running the following command:
+
+```
+npm run generate:ui
+```
+
+## Use Case
+
+AI-powered document generator that can help you generate documents with a chat interface and simple markdown editor.
+
+To update the workflow, you can modify the code in [`workflow.ts`](app/workflow.ts).
+
+You can start by sending an request on the [chat UI](http://localhost:3000) or you can test the `/api/chat` endpoint with the following curl request:
+
+```shell
+curl --location 'localhost:3000/api/chat' \
+--header 'Content-Type: application/json' \
+--data '{ "messages": [{ "role": "user", "content": "Compare the financial performance of Apple and Tesla" }] }'
+```
+
+## Learn More
+
+To learn more about LlamaIndex, take a look at the following resources:
+
+- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex (Python features).
+- [LlamaIndexTS Documentation](https://ts.llamaindex.ai/docs/llamaindex) - learn about LlamaIndex (Typescript features).
+- [Workflows Introduction](https://ts.llamaindex.ai/docs/llamaindex/modules/workflows) - learn about LlamaIndexTS workflows.
+
+You can check out [the LlamaIndexTS GitHub repository](https://github.com/run-llama/LlamaIndexTS) - your feedback and contributions are welcome!
@@ -1,5 +1,5 @@
 import { extractLastArtifact } from "@llamaindex/server";
-import { ChatMemoryBuffer, LLM, Settings } from "llamaindex";
+import { ChatMemoryBuffer, MessageContent, Settings } from "llamaindex";

 import {
  agentStreamEvent,
@@ -40,7 +40,7 @@ export const UIEventSchema = z.object({
 export type UIEvent = z.infer<typeof UIEventSchema>;

 const planEvent = workflowEvent<{
-  userInput: string;
+  userInput: MessageContent;
  context?: string | undefined;
 }>();

@@ -68,32 +68,28 @@ const artifactEvent = workflowEvent<{
  };
 }>();

-export function createDocumentArtifactWorkflow(reqBody: any, llm?: LLM) {
-  if (!llm) {
-    llm = Settings.llm;
-  }
+export function workflowFactory(reqBody: any) {
+  const llm = Settings.llm;
+
  const { withState, getContext } = createStatefulMiddleware(() => {
    return {
-      memory: new ChatMemoryBuffer({
-        llm,
-        chatHistory: reqBody.chatHistory,
-      }),
+      memory: new ChatMemoryBuffer({ llm }),
      lastArtifact: extractLastArtifact(reqBody),
    };
  });
  const workflow = withState(createWorkflow());

-  workflow.handle([startAgentEvent], async ({ data: { userInput } }) => {
+  workflow.handle([startAgentEvent], async ({ data }) => {
+    const { userInput, chatHistory = [] } = data;
    // Prepare chat history
    const { state } = getContext();
    // Put user input to the memory
    if (!userInput) {
      throw new Error("Missing user input to start the workflow");
    }
-    state.memory.put({
-      role: "user",
-      content: userInput,
-    });
+    state.memory.set(chatHistory);
+    state.memory.put({ role: "user", content: userInput });
+
    return planEvent.with({
      userInput,
      context: state.lastArtifact
@@ -175,15 +171,6 @@ export function createDocumentArtifactWorkflow(reqBody: any, llm?: LLM) {
      role: "assistant",
      content: `Planning for the document generation: \n${response.text}`,
    });
-    sendEvent(
-      uiEvent.with({
-        type: "ui_event",
-        data: {
-          state: "generate",
-          requirement: requirement.requirement,
-        },
-      }),
-    );
    return generateArtifactEvent.with({
      requirement,
    });
@@ -12,11 +12,12 @@ from llama_index.server.services.llamacloud.generate import (
    load_to_llamacloud,
 )

+
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger()


-def generate_datasource():
+def generate_index():
    init_settings()
    logger.info("Generate index for the provided data")

@@ -27,5 +28,26 @@ def generate_datasource():
    load_to_llamacloud(index, logger=logger)


-if __name__ == "__main__":
-    generate_datasource()
+def generate_ui_for_workflow():
+    """
+    Generate UI for UIEventData event in app/workflow.py
+    """
+    import asyncio
+    from llama_index.llms.openai import OpenAI
+    from main import COMPONENT_DIR
+
+    # To generate UI components for additional event types,
+    # import the corresponding data model (e.g., MyCustomEventData)
+    # and run the generate_ui_for_workflow function with the imported model.
+    # Make sure the output filename of the generated UI component matches the event type (here `ui_event`)
+    try:
+        from app.workflow import UIEventData  # type: ignore
+    except ImportError:
+        raise ImportError("Couldn't generate UI component for the current workflow.")
+    from llama_index.server.gen_ui import generate_event_component
+
+    # works also well with Claude 3.7 Sonnet or Gemini Pro 2.5
+    llm = OpenAI(model="gpt-4.1")
+    code = asyncio.run(generate_event_component(event_cls=UIEventData, llm=llm))
+    with open(f"{COMPONENT_DIR}/ui_event.jsx", "w") as f:
+        f.write(code)
@@ -1,7 +1,8 @@
+import { generateEventComponent } from "@llamaindex/server";
 import * as dotenv from "dotenv";
 import "dotenv/config";
 import * as fs from "fs/promises";
-import { LLamaCloudFileService } from "llamaindex";
+import { LLamaCloudFileService, OpenAI } from "llamaindex";
 import * as path from "path";
 import { getIndex } from "./app/data";
 import { initSettings } from "./app/settings";
@@ -88,7 +89,7 @@ async function loadAndIndex() {
  console.log(`Successfully uploaded documents to LlamaCloud!`);
 }

-(async () => {
+async function generateDatasource() {
  try {
    checkRequiredEnvVars();
    initSettings();
@@ -97,4 +98,39 @@ async function loadAndIndex() {
  } catch (error) {
    console.error("Error generating storage.", error);
  }
+}
+
+async function generateUi() {
+  // Also works well with Claude 3.5 Sonnet and Google Gemini 2.5 Pro
+  const llm = new OpenAI({ model: "gpt-4.1" });
+
+  const workflowModule = await import("./app/workflow");
+  const UIEventSchema = (workflowModule as any).UIEventSchema;
+  if (!UIEventSchema) {
+    throw new Error(
+      "To generate the UI, you must define a UIEventSchema in your workflow.",
+    );
+  }
+
+  const generatedCode = await generateEventComponent(UIEventSchema, llm);
+  // Write the generated code to components/ui_event.ts
+  await fs.writeFile("components/ui_event.jsx", generatedCode);
+}
+
+(async () => {
+  const args = process.argv.slice(2);
+  const command = args[0];
+
+  initSettings();
+
+  if (command === "datasource") {
+    await generateDatasource();
+  } else if (command === "ui") {
+    await generateUi();
+  } else {
+    console.error(
+      'Invalid command. Please use "datasource" or "ui". Running "datasource" by default.',
+    );
+    await generateDatasource(); // Default behavior or could throw an error
+  }
 })();
@@ -1,8 +1,12 @@
+import os
+
 from llama_index.core import Settings
 from llama_index.embeddings.openai import OpenAIEmbedding
 from llama_index.llms.openai import OpenAI


 def init_settings():
+    if os.getenv("OPENAI_API_KEY") is None:
+        raise RuntimeError("OPENAI_API_KEY is missing in environment variables")
    Settings.llm = OpenAI(model="gpt-4o-mini")
    Settings.embed_model = OpenAIEmbedding(model="text-embedding-3-small")
@@ -17,8 +17,10 @@ def create_app():
        ui_config=UIConfig(
            component_dir=COMPONENT_DIR,
            app_title="Chat App",
+            dev_mode=True,  # Please disable this in production
        ),
        logger=logger,
+        env="dev",
    )
    # You can also add custom FastAPI routes to app
    app.add_api_route("/api/health", lambda: {"message": "OK"}, status_code=200)
@@ -12,7 +12,7 @@ dependencies = [
    "pydantic<2.10",
    "aiostream>=0.5.2,<0.6.0",
    "llama-index-core>=0.12.28,<0.13.0",
-    "llama-index-server>=0.1.15,<0.2.0",
+    "llama-index-server>=0.1.16,<0.2.0",
 ]

 [project.optional-dependencies]
@@ -0,0 +1,6 @@
+{
+  "watch": ["src/**/*.ts"],
+  "exec": "nodemon --exec tsx src/index.ts",
+  "ext": "js ts",
+  "ignore": ["src/app/workflow_*.ts"]
+}
@@ -5,21 +5,22 @@
    "generate": "tsx src/generate.ts datasource",
    "generate:datasource": "tsx src/generate.ts datasource",
    "generate:ui": "tsx src/generate.ts ui",
-    "dev": "tsx watch src/index.ts",
+    "dev": "nodemon",
    "start": "tsx src/index.ts"
  },
  "dependencies": {
-    "@llamaindex/openai": "0.2.0",
-    "@llamaindex/server": "0.2.0",
-    "@llamaindex/workflow": "1.1.1",
-    "@llamaindex/tools": "0.0.9",
+    "@llamaindex/openai": "~0.4.0",
+    "@llamaindex/server": "~0.2.1",
+    "@llamaindex/workflow": "~1.1.3",
+    "@llamaindex/tools": "~0.0.11",
+    "llamaindex": "~0.11.0",
    "dotenv": "^16.4.7",
-    "zod": "^3.23.8",
-    "llamaindex": "0.10.5"
+    "zod": "^3.23.8"
  },
  "devDependencies": {
    "@types/node": "^20.10.3",
    "tsx": "^4.7.2",
-    "typescript": "^5.3.2"
+    "typescript": "^5.3.2",
+    "nodemon": "^3.1.10"
  }
 }
@@ -3,7 +3,7 @@ import { Settings } from "llamaindex";

 export function initSettings() {
  Settings.llm = new OpenAI({
-    model: "gpt-4o-mini",
+    model: "gpt-4.1",
  });
  Settings.embedModel = new OpenAIEmbedding({
    model: "text-embedding-3-small",
@@ -1,14 +1,10 @@
 import "dotenv/config";
 import { SimpleDirectoryReader } from "@llamaindex/readers/directory";
-import {
-  OpenAI,
-  storageContextFromDefaults,
-  VectorStoreIndex,
-} from "llamaindex";
+import { storageContextFromDefaults, VectorStoreIndex } from "llamaindex";
 import { initSettings } from "./app/settings";
 import fs from "fs";
 import { generateEventComponent } from "@llamaindex/server";
-import { UIEventSchema } from "./app/workflow";
+import { OpenAI } from "@llamaindex/openai";

 async function generateDatasource() {
  console.log(`Generating storage context...`);
@@ -30,6 +26,14 @@ async function generateUi() {
  // Also works well with Claude 3.5 Sonnet and Google Gemini 2.5 Pro
  const llm = new OpenAI({ model: "gpt-4.1" });

+  const workflowModule = await import("./app/workflow");
+  const UIEventSchema = (workflowModule as any).UIEventSchema;
+  if (!UIEventSchema) {
+    throw new Error(
+      "To generate the UI, you must define a UIEventSchema in your workflow.",
+    );
+  }
+
  // You can also generate for other workflow events
  const generatedCode = await generateEventComponent(UIEventSchema, llm);
  // Write the generated code to components/ui_event.ts
@@ -10,5 +10,6 @@ new LlamaIndexServer({
  uiConfig: {
    appTitle: "LlamaIndex App",
    componentsDir: "components",
+    devMode: true,
  },
 }).start();
@@ -1 +1,5 @@
-server/
+# server contains Nextjs frontend code (not compiled)
+server/
+
+# temp is the copy of next folder but without API folder, used to build frontend static files
+temp/
@@ -1,5 +1,18 @@
 # @llamaindex/server

+## 0.2.2
+
+### Patch Changes
+
+- 25fba43: refactor: migrate to Nextjs Route Handler
+- 6f75d4a: fix: unsupported language in code gen workflow
+
+## 0.2.1
+
+### Patch Changes
+
+- f072308: feat: add dev mode UI
+
 ## 0.2.0

 ### Minor Changes
@@ -4,10 +4,10 @@ LlamaIndexServer is a Next.js-based application that allows you to quickly launc

 ## Features

- Serving a workflow as a chatbot
+- Add a sophisticated chatbot UI to your LlamaIndex workflow
+- Edit code and document artifacts in an OpenAI Canvas-style UI
+- Extendable UI components for events and headers
 - Built on Next.js for high performance and easy API development
- Optional built-in chat UI with extendable UI components
- Prebuilt development code

 ## Installation

@@ -21,9 +21,11 @@ Create an `index.ts` file and add the following code:

 ```ts
 import { LlamaIndexServer } from "@llamaindex/server";
+import { openai } from "@llamaindex/openai";
+import { agent } from "@llamaindex/workflow";
 import { wiki } from "@llamaindex/tools"; // or any other tool

-const createWorkflow = () => agent({ tools: [wiki()] });
+const createWorkflow = () => agent({ tools: [wiki()], llm: openai("gpt-4o") });

 new LlamaIndexServer({
  workflow: createWorkflow,
@@ -34,6 +36,8 @@ new LlamaIndexServer({
 }).start();
 ```

+The `createWorkflow` function is a factory function that creates an [Agent Workflow](https://ts.llamaindex.ai/docs/llamaindex/modules/agents/agent_workflow) with a tool that retrieves information from Wikipedia in this case. For more details, read about the [Workflow factory contract](#workflow-factory-contract).
+
 ## Running the Server

 In the same directory as `index.ts`, run the following command to start the server:
@@ -54,16 +58,75 @@ curl -X POST "http://localhost:3000/api/chat" -H "Content-Type: application/json

 The `LlamaIndexServer` accepts the following configuration options:

- `workflow`: A callable function that creates a workflow instance for each request
+- `workflow`: A callable function that creates a workflow instance for each request. See [Workflow factory contract](#workflow-factory-contract) for more details.
 - `uiConfig`: An object to configure the chat UI containing the following properties:
  - `appTitle`: The title of the application (default: `"LlamaIndex App"`)
  - `starterQuestions`: List of starter questions for the chat UI (default: `[]`)
  - `componentsDir`: The directory for custom UI components rendering events emitted by the workflow. The default is undefined, which does not render custom UI components.
  - `llamaCloudIndexSelector`: Whether to show the LlamaCloud index selector in the chat UI (requires `LLAMA_CLOUD_API_KEY` to be set in the environment variables) (default: `false`)
+  - `dev_mode`: When enabled, you can update workflow code in the UI and see the changes immediately. It's currently in beta and only supports updating workflow code at `app/src/workflow.ts`. Please start server in dev mode (`npm run dev`) to use see this reload feature enabled.

 LlamaIndexServer accepts all the configuration options from Nextjs Custom Server such as `port`, `hostname`, `dev`, etc.
 See all Nextjs Custom Server options [here](https://nextjs.org/docs/app/building-your-application/configuring/custom-server).

+## Workflow factory contract
+
+The `workflow` provided will be called for each chat request to initialize a new workflow instance. The contract of the generated workflow must be the same as for the [Agent Workflow](https://ts.llamaindex.ai/docs/llamaindex/modules/agents/agent_workflow).
+
+This means that the workflow must handle a `startAgentEvent` event, which is the entry point of the workflow and contains the following information in it's `data` property:
+
+```typescript
+{
+  userInput: MessageContent;
+  chatHistory?: ChatMessage[] | undefined;
+};
+```
+
+The `userInput` is the latest user message and the `chatHistory` is the list of messages exchanged between the user and the workflow so far.
+
+Furthermore, the workflow must stop with a `stopAgentEvent` event to mark the end of the workflow. In between, the workflow can emit [UI events](##AI-generated-UI-Components) to render custom UI components and [Artifact events](##Sending-Artifacts-to-the-UI) to send structured data like generated documents or code snippets to the UI.
+
+```ts
+import {
+  createStatefulMiddleware,
+  createWorkflow,
+  startAgentEvent,
+} from "@llamaindex/workflow";
+import { ChatMemoryBuffer, type ChatMessage, Settings } from "llamaindex";
+import { openai } from "@llamaindex/openai";
+import { wiki } from "@llamaindex/tools";
+
+Settings.llm = openai("gpt-4o");
+
+export const workflowFactory = async () => {
+  const workflow = createWorkflow();
+
+  workflow.handle([startAgentEvent], async ({ data }) => {
+    const { state, sendEvent } = getContext();
+    const messages = data.chatHistory;
+
+    const toolCallResponse = await chatWithTools(
+      Settings.llm,
+      [wiki()],
+      messages,
+    );
+
+    // using result from tool call and use `sendEvent` to emit the next event...
+  });
+
+  // define more workflow handling logic here...
+
+  // Finally stop with a `stopAgentEvent` event to mark the end of the workflow.
+  // return stopAgentEvent.with({
+  //   result: "This is the end!",
+  // });
+
+  return workflow;
+};
+```
+
+To generate sophisticated examples of workflows, you best use the [create-llama](https://github.com/run-llama/create-llama) project.
+
 ## AI-generated UI Components

 The LlamaIndex server provides support for rendering workflow events using custom UI components, allowing you to extend and customize the chat interface.
@@ -137,6 +200,66 @@ new LlamaIndexServer({
 }).start();
 ```

+## Sending Artifacts to the UI
+
+In addition to UI events for custom components, LlamaIndex Server supports a special `ArtifactEvent` to send structured data like generated documents or code snippets to the UI. These artifacts are displayed in a dedicated "Canvas" panel in the chat interface.
+
+### Artifact Event Structure
+
+To send an artifact, your workflow needs to emit an event with `type: "artifact"`. The `data` payload of this event should include:
+
+- `type`: A string indicating the type of artifact (e.g., `"document"`, `"code"`).
+- `created_at`: A timestamp (e.g., `Date.now()`) indicating when the artifact was created.
+- `data`: An object containing the specific details of the artifact. The structure of this object depends on the artifact `type`.
+
+### Defining and Sending an ArtifactEvent
+
+First, define your artifact event using `workflowEvent` from `@llamaindex/workflow`:
+
+```typescript
+import { workflowEvent } from "@llamaindex/workflow";
+
+// Example for a document artifact
+const artifactEvent = workflowEvent<{
+  type: "artifact"; // Must be "artifact"
+  data: {
+    type: "document"; // Custom type for your artifact (e.g., "document", "code")
+    created_at: number;
+    data: {
+      // Specific data for the document artifact type
+      title: string;
+      content: string;
+      type: "markdown" | "html"; // document format
+    };
+  };
+}>();
+```
+
+Then, within your workflow logic, use `sendEvent` (obtained from `getContext()`) to emit the event:
+
+```typescript
+// Assuming 'sendEvent' is available in your workflow handler
+// and 'documentDetails' contains the content for the artifact.
+
+sendEvent(
+  artifactEvent.with({
+    type: "artifact", // This top-level type must be "artifact"
+    data: {
+      type: "document", // This is your specific artifact type
+      created_at: Date.now(),
+      data: {
+        title: "My Generated Document",
+        content: "# Hello World
+This is a markdown document.",
+        type: "markdown",
+      },
+    },
+  }),
+);
+```
+
+This will send the artifact to the LlamaIndex Server UI, where it will be rendered in the [ChatCanvasPanel](/packages/server/next/app/components/ui/chat/canvas/panel.tsx) by a renderer depending on the artifact type. For type `document` this is using the [DocumentArtifactViewer](https://github.com/run-llama/chat-ui/blob/bacb75fc6edceacf742fba18632404a2483b5a81/packages/chat-ui/src/chat/canvas/artifacts/document.tsx#L17).
+
 ## Default Endpoints and Features

 ### Chat Endpoint
@@ -6,7 +6,7 @@ This directory contains examples of how to use the LlamaIndex Server.

 ```bash
 export OPENAI_API_KEY=your_openai_api_key
-npx tsx simple-workflow/calculator.ts
+pnpm run dev
 ```

 ## Open browser at http://localhost:3000
@@ -39,5 +39,5 @@ new LlamaIndexServer({
    appTitle: "LlamaIndex App",
    starterQuestions: ["What is the color of the dog?"],
  },
-  port: 4100,
+  port: 3000,
 }).start();
@@ -0,0 +1,21 @@
+This example shows how to use the dev mode of the server.
+
+First, we need to set `devMode` to `true` in the `uiConfig` of the server.
+
+```ts
+new LlamaIndexServer({
+  workflow: workflowFactory,
+  uiConfig: {
+    appTitle: "Calculator",
+    devMode: true,
+  },
+  port: 3000,
+}).start();
+```
+
+Export OpenAI API key and start the server in dev mode.
+
+```bash
+export OPENAI_API_KEY=<your-openai-api-key>
+npx nodemon --exec tsx index.ts --ignore src/app/workflow_*.ts
+```
@@ -0,0 +1,15 @@
+import { LlamaIndexServer } from "@llamaindex/server";
+import { workflowFactory } from "./src/app/workflow";
+
+new LlamaIndexServer({
+  workflow: workflowFactory,
+  uiConfig: {
+    appTitle: "Calculator",
+    devMode: true,
+    starterQuestions: [
+      "What is the weather in Tokyo?",
+      "What is the weather in New York?",
+    ],
+  },
+  port: 3000,
+}).start();
@@ -0,0 +1,16 @@
+import { agent } from "@llamaindex/workflow";
+import { tool } from "llamaindex";
+import { z } from "zod";
+
+export const workflowFactory = async () => {
+  return agent({
+    tools: [
+      tool({
+        name: "weather",
+        description: "Get the weather in a specific city",
+        parameters: z.object({ city: z.string() }),
+        execute: ({ city }) => `The weather in ${city} is sunny`,
+      }),
+    ],
+  });
+};
@@ -4,7 +4,7 @@
  "private": true,
  "scripts": {
    "typecheck": "tsc --noEmit",
-    "dev": "tsx simple-workflow/calculator.ts"
+    "dev": "nodemon --exec tsx simple-workflow/calculator.ts"
  },
  "dependencies": {
    "@llamaindex/openai": "^0.2.0",
@@ -18,6 +18,7 @@
  },
  "devDependencies": {
    "@types/node": "^20.10.3",
+    "nodemon": "^3.1.10",
    "tsx": "^4.7.2",
    "typescript": "^5.3.2"
  }
@@ -20,5 +20,5 @@ new LlamaIndexServer({
    appTitle: "Calculator",
    starterQuestions: ["1 + 1", "2 + 2"],
  },
-  port: 4000,
+  port: 3000,
 }).start();
@@ -0,0 +1,32 @@
+import { getEnv } from "@llamaindex/env";
+import { LLamaCloudFileService } from "llamaindex";
+import { NextRequest, NextResponse } from "next/server";
+
+export async function GET(request: NextRequest): Promise<NextResponse> {
+  if (!getEnv("LLAMA_CLOUD_API_KEY")) {
+    return NextResponse.json(
+      {
+        error: "env variable LLAMA_CLOUD_API_KEY is required to use LlamaCloud",
+      },
+      { status: 500 },
+    );
+  }
+
+  try {
+    const config = {
+      projects: await LLamaCloudFileService.getAllProjectsWithPipelines(),
+      pipeline: {
+        pipeline: getEnv("LLAMA_CLOUD_INDEX_NAME"),
+        project: getEnv("LLAMA_CLOUD_PROJECT_NAME"),
+      },
+    };
+    return NextResponse.json(config, { status: 200 });
+  } catch (error) {
+    return NextResponse.json(
+      {
+        error: "Failed to fetch LlamaCloud configuration",
+      },
+      { status: 500 },
+    );
+  }
+}
@@ -0,0 +1,77 @@
+import { type AgentInputData } from "@llamaindex/workflow";
+import { type Message } from "ai";
+import { type MessageType } from "llamaindex";
+import { NextRequest, NextResponse } from "next/server";
+
+// import chat utils
+import { toDataStream } from "./utils/stream";
+import { sendSuggestedQuestionsEvent } from "./utils/suggestion";
+import { runWorkflow } from "./utils/workflow";
+
+// import workflow factory from local file
+import { workflowFactory } from "../../../../app/workflow";
+
+export async function POST(req: NextRequest) {
+  try {
+    const reqBody = await req.json();
+    const { messages } = reqBody as { messages: Message[] };
+    const chatHistory = messages.map((message) => ({
+      role: message.role as MessageType,
+      content: message.content,
+    }));
+
+    const lastMessage = messages[messages.length - 1];
+    if (lastMessage?.role !== "user") {
+      return NextResponse.json(
+        {
+          detail: "Messages cannot be empty and last message must be from user",
+        },
+        { status: 400 },
+      );
+    }
+    const workflowInput: AgentInputData = {
+      userInput: lastMessage.content,
+      chatHistory,
+    };
+
+    const abortController = new AbortController();
+    req.signal.addEventListener("abort", () =>
+      abortController.abort("Connection closed"),
+    );
+
+    const workflow = await workflowFactory(reqBody);
+    const workflowEventStream = await runWorkflow(
+      workflow,
+      workflowInput,
+      abortController.signal,
+    );
+
+    const dataStream = toDataStream(workflowEventStream, {
+      callbacks: {
+        onFinal: async (completion, dataStreamWriter) => {
+          chatHistory.push({
+            role: "assistant" as MessageType,
+            content: completion,
+          });
+          await sendSuggestedQuestionsEvent(dataStreamWriter, chatHistory);
+        },
+      },
+    });
+
+    return new Response(dataStream, {
+      status: 200,
+      headers: {
+        "Content-Type": "text/plain; charset=utf-8",
+        "X-Vercel-AI-Data-Stream": "v1",
+      },
+    });
+  } catch (error) {
+    console.error("Chat handler error:", error);
+    return NextResponse.json(
+      {
+        detail: (error as Error).message || "Internal server error",
+      },
+      { status: 500 },
+    );
+  }
+}
@@ -1,20 +1,19 @@
 import fs from "fs";
-import type { IncomingMessage, ServerResponse } from "http";
+import { NextRequest, NextResponse } from "next/server";
 import path from "path";
 import { promisify } from "util";
-import { sendJSONResponse } from "../utils/request";

-export const getComponents = async (
-  req: IncomingMessage,
-  res: ServerResponse,
-  componentsDir: string,
-) => {
+export async function GET(request: NextRequest) {
+  const params = request.nextUrl.searchParams;
+  const componentsDir = params.get("componentsDir") || "components";
+
  try {
    const exists = await promisify(fs.exists)(componentsDir);
    if (!exists) {
-      return sendJSONResponse(res, 404, {
-        error: "Components directory not found",
-      });
+      return NextResponse.json(
+        { error: "Components directory not found" },
+        { status: 404 },
+      );
    }

    const files = await promisify(fs.readdir)(componentsDir);
@@ -40,12 +39,15 @@ export const getComponents = async (
      }),
    );

-    sendJSONResponse(res, 200, components);
+    return NextResponse.json(components, { status: 200 });
  } catch (error) {
    console.error("Error reading components:", error);
-    sendJSONResponse(res, 500, { error: "Failed to read components" });
+    return NextResponse.json(
+      { error: "Failed to read components" },
+      { status: 500 },
+    );
  }
-};
+}

 function filterDuplicateComponents(files: string[]) {
  const compMap = new Map<string, string>();
@@ -0,0 +1,96 @@
+import { exec } from "child_process";
+import fs from "fs";
+import { NextRequest, NextResponse } from "next/server";
+import path from "path";
+import { promisify } from "util";
+
+const DEFAULT_WORKFLOW_FILE_PATH = "src/app/workflow.ts"; // TODO: we can make it as a parameter in server later
+
+export async function GET(request: NextRequest) {
+  const filePath = DEFAULT_WORKFLOW_FILE_PATH;
+
+  const fileExists = await promisify(fs.exists)(DEFAULT_WORKFLOW_FILE_PATH);
+  if (!fileExists) {
+    return NextResponse.json(
+      {
+        detail: `Dev mode is currently in beta. It only supports updating workflow file at ${filePath}`,
+      },
+      { status: 404 },
+    );
+  }
+
+  const content = await promisify(fs.readFile)(filePath, "utf-8");
+  const last_modified = fs.statSync(filePath).mtime.getTime();
+
+  return NextResponse.json(
+    { content, file_path: filePath, last_modified },
+    { status: 200 },
+  );
+}
+
+export async function PUT(request: NextRequest) {
+  const filePath = DEFAULT_WORKFLOW_FILE_PATH;
+  const { content } = await request.json();
+
+  const fileExists = await promisify(fs.exists)(filePath);
+  if (!fileExists) {
+    return NextResponse.json(
+      {
+        detail: `Dev mode is currently in beta. It only supports updating workflow file at ${DEFAULT_WORKFLOW_FILE_PATH}`,
+      },
+      { status: 404 },
+    );
+  }
+
+  try {
+    const resolvedFilePath = path.resolve(DEFAULT_WORKFLOW_FILE_PATH);
+    const result = await validateTypeScriptFile(resolvedFilePath, content);
+
+    if (!result.isValid) {
+      return NextResponse.json(
+        {
+          detail: result.errors.join("\n"),
+        },
+        { status: 400 },
+      );
+    }
+
+    await promisify(fs.writeFile)(filePath, content);
+    return NextResponse.json({ content }, { status: 200 });
+  } catch (error) {
+    console.error("Error updating workflow file:", error);
+    return NextResponse.json(
+      { error: "Failed to update workflow file" },
+      { status: 500 },
+    );
+  }
+}
+
+// use typescript package to validate the file syntax and imports
+async function validateTypeScriptFile(filePath: string, content: string) {
+  // Update workflow file directly will cause the server restart immediately.
+  // So we create a temporary file with the same content in the same directory as the workflow file
+  // This file will be used to validate the file syntax and imports. It will be deleted after validation.
+  const tempFilePath = path.join(
+    path.dirname(filePath),
+    `workflow_${Date.now()}.ts`,
+  );
+  fs.writeFileSync(tempFilePath, content);
+
+  const errors = [];
+  try {
+    const tscCommand = `npx tsc ${tempFilePath} --noEmit --skipLibCheck true`;
+    await promisify(exec)(tscCommand);
+  } catch (error) {
+    const errorMessage = (error as { stdout: string })?.stdout;
+    errors.push(errorMessage);
+  } finally {
+    // Clean up temporary file
+    if (fs.existsSync(tempFilePath)) fs.unlinkSync(tempFilePath);
+  }
+
+  return {
+    isValid: errors.length === 0,
+    errors: errors,
+  };
+}
@@ -0,0 +1,24 @@
+import fs from "fs";
+import { NextRequest, NextResponse } from "next/server";
+import { promisify } from "util";
+
+export async function GET(
+  request: NextRequest,
+  { params }: { params: Promise<{ slug: string[] }> },
+) {
+  const filePath = (await params).slug.join("/");
+
+  if (!filePath.startsWith("output") && !filePath.startsWith("data")) {
+    return NextResponse.json({ error: "No permission" }, { status: 400 });
+  }
+
+  const decodedFilePath = decodeURIComponent(filePath);
+  const fileExists = await promisify(fs.exists)(decodedFilePath);
+
+  if (fileExists) {
+    const fileBuffer = await promisify(fs.readFile)(decodedFilePath);
+    return new NextResponse(fileBuffer);
+  } else {
+    return NextResponse.json({ error: "File not found" }, { status: 404 });
+  }
+}
@@ -13,6 +13,7 @@ import CustomChatMessages from "./chat-messages";
 import { DynamicEventsErrors } from "./custom/events/dynamic-events-errors";
 import { fetchComponentDefinitions } from "./custom/events/loader";
 import { ComponentDef } from "./custom/events/types";
+import { DevModePanel } from "./dev-mode-panel";

 export default function ChatSection() {
  const handler = useChat({
@@ -35,12 +36,13 @@ export default function ChatSection() {
        <ChatHeader />
        <ChatUI
          handler={handler}
-          className="flex min-h-0 flex-1 flex-row justify-center gap-4 px-4 py-0"
+          className="relative flex min-h-0 flex-1 flex-row justify-center gap-4 px-4 py-0"
        >
          <ResizablePanelGroup direction="horizontal">
            <ChatSectionPanel />
            <ChatCanvasPanel />
          </ResizablePanelGroup>
+          <DevModePanel />
        </ChatUI>
      </div>
      <ChatInjection />
@@ -0,0 +1,273 @@
+"use client";
+
+import {
+  CodeEditor,
+  fileExtensionToEditorLang,
+} from "@llamaindex/chat-ui/widgets";
+import { AlertCircle, Loader2 } from "lucide-react";
+import { useEffect, useMemo, useState } from "react";
+import { Button } from "../button";
+import { getConfig } from "../lib/utils";
+
+const API_PATH = "/api/dev/files/workflow";
+const POLLING_TIMEOUT = 30_000; // 30 seconds
+
+type WorkflowFile = {
+  last_modified: number;
+  file_path: string;
+  content: string;
+};
+
+export function DevModePanel() {
+  const devModeEnabled = getConfig("DEV_MODE");
+  if (!devModeEnabled) return null;
+  return <DevModePanelComp />;
+}
+
+function DevModePanelComp() {
+  const [devModeOpen, setDevModeOpen] = useState(false);
+
+  const [isFetching, setIsFetching] = useState(false);
+  const [fetchingError, setFetchingError] = useState<string | null>();
+  const [workflowFile, setWorkflowFile] = useState<WorkflowFile | null>(null);
+
+  const [updatedCode, setUpdatedCode] = useState<string | null>(null);
+  const [isSaving, setIsSaving] = useState(false);
+  const [saveError, setSaveError] = useState<string | null>(null);
+
+  const [isPolling, setIsPolling] = useState(false);
+  const [pollingError, setPollingError] = useState<string | null>(null);
+
+  async function fetchWorkflowCode() {
+    try {
+      setIsFetching(true);
+      const response = await fetch(API_PATH);
+      const data = await response.json();
+
+      if (!response.ok) {
+        throw new Error(data?.detail ?? "Unknown error");
+      }
+
+      setWorkflowFile(data);
+      setFetchingError(null);
+    } catch (error) {
+      const errorMessage =
+        error instanceof Error ? error.message : "Unknown error";
+      setFetchingError(errorMessage);
+      console.warn("Error fetching workflow code:", error);
+    } finally {
+      setIsFetching(false);
+    }
+  }
+
+  async function restartingWorkflow() {
+    if (!workflowFile) return;
+
+    const initialLastModified = workflowFile.last_modified;
+    setIsPolling(true);
+    setPollingError(null);
+
+    const pollStartTime = Date.now();
+
+    // interval refetching the updated workflow code
+    const poll = async () => {
+      if (Date.now() - pollStartTime > POLLING_TIMEOUT) {
+        setPollingError(
+          `Server not responding after ${POLLING_TIMEOUT / 1000} seconds.`,
+        );
+        return;
+      }
+
+      try {
+        const pollResponse = await fetch(API_PATH);
+        const pollData = (await pollResponse.json()) as WorkflowFile;
+        if (pollData.last_modified !== initialLastModified) {
+          setWorkflowFile(pollData);
+          setUpdatedCode(pollData.content);
+          setIsPolling(false);
+          setPollingError(null);
+          setDevModeOpen(false);
+        } else {
+          setTimeout(poll, 2000);
+        }
+      } catch (error) {
+        console.info("Polling error", error);
+        setTimeout(poll, 2000);
+      }
+    };
+
+    setTimeout(poll, 2000);
+  }
+
+  const handleResetCode = () => {
+    setUpdatedCode(workflowFile?.content ?? null);
+    setSaveError(null);
+  };
+
+  const handleSaveCode = async () => {
+    if (!workflowFile) return;
+
+    try {
+      setIsSaving(true);
+      const response = await fetch(API_PATH, {
+        method: "PUT",
+        headers: {
+          Accept: "application/json",
+          "Content-Type": "application/json",
+        },
+        body: JSON.stringify({
+          content: updatedCode,
+          file_path: workflowFile.file_path,
+        }),
+      });
+      const data = await response.json();
+      if (!response.ok) {
+        throw new Error(data?.detail ?? "Unknown error");
+      }
+      setSaveError(null);
+      await restartingWorkflow();
+    } catch (error) {
+      console.warn("Error saving workflow code:", error);
+      setSaveError(
+        error instanceof Error
+          ? error.message
+          : "Unknown error happened when saving workflow code",
+      );
+    } finally {
+      setIsSaving(false);
+    }
+  };
+
+  useEffect(() => {
+    if (devModeOpen) {
+      fetchWorkflowCode();
+    }
+  }, [devModeOpen]);
+
+  const codeEditorLanguage = useMemo(() => {
+    if (!workflowFile?.file_path) return undefined;
+    return fileExtensionToEditorLang(
+      workflowFile.file_path.split(".").pop() ?? "",
+    );
+  }, [workflowFile]);
+
+  return (
+    <>
+      <Button
+        onClick={() => setDevModeOpen(!devModeOpen)}
+        className="fixed right-2 top-1/2 origin-right -translate-y-1/2 rotate-90 transform rounded-l-md shadow-md transition-transform hover:-translate-x-1"
+      >
+        Dev Mode
+      </Button>
+
+      {isPolling && (
+        <div className="fixed inset-0 z-50 flex flex-col items-center justify-center bg-black/50 backdrop-blur-sm">
+          {!pollingError && (
+            <>
+              <Loader2 className="mb-4 h-16 w-16 animate-spin text-white" />
+              <p className="text-lg font-semibold text-white">
+                Applying changes and restarting server...
+              </p>
+              <p className="mt-2 text-sm text-slate-300">
+                Please wait for a while then you can start chatting with the
+                updated workflow.
+              </p>
+            </>
+          )}
+          {pollingError && (
+            <div className="bg-destructive/20 text-destructive-foreground mt-4 max-w-md rounded-md p-4 text-center">
+              <div className="mb-2 flex items-center justify-center gap-2">
+                <AlertCircle className="shrink-0" size={16} />
+                <h6 className="text-sm font-medium">Server Starting Error</h6>
+              </div>
+              <p className="text-sm">{pollingError}</p>
+
+              <p className="text-sm">
+                Please reload the page and check server logs.
+              </p>
+            </div>
+          )}
+        </div>
+      )}
+
+      <div
+        className={`border-border fixed right-0 top-0 z-10 h-full w-full border-l shadow-xl transition-all duration-300 ease-in-out ${
+          devModeOpen ? "translate-x-0 bg-black/50" : "translate-x-full"
+        }`}
+        onClick={() => setDevModeOpen(false)}
+      >
+        <div
+          className={`bg-background ml-auto flex h-full w-[800px] flex-col p-4`}
+          onClick={(e) => e.stopPropagation()}
+        >
+          <div className="mb-4 flex items-center justify-between">
+            <div>
+              <h2 className="text-xl font-bold">Workflow Editor</h2>
+              <p className="text-muted-foreground text-sm">
+                {isFetching ? (
+                  "Loading..."
+                ) : workflowFile ? (
+                  <>
+                    Edit the code of <b>{workflowFile.file_path}</b> and save to
+                    apply changes to your workflow.
+                  </>
+                ) : (
+                  ""
+                )}
+              </p>
+            </div>
+            <Button
+              variant="outline"
+              size="sm"
+              onClick={() => setDevModeOpen(false)}
+            >
+              Close
+            </Button>
+          </div>
+          <div className="flex-1 overflow-auto">
+            {fetchingError ? (
+              <div className="bg-destructive/10 text-destructive/70 mb-4 flex items-center gap-2 rounded-md p-4">
+                <AlertCircle className="shrink-0" size={16} />
+                <p className="text-sm font-medium">{fetchingError}</p>
+              </div>
+            ) : (
+              <CodeEditor
+                code={updatedCode ?? workflowFile?.content ?? ""}
+                onChange={setUpdatedCode}
+                language={codeEditorLanguage}
+              />
+            )}
+          </div>
+          <div className="mt-4 flex flex-col">
+            {saveError && (
+              <div className="bg-destructive/10 text-destructive/70 mb-4 rounded-md p-4">
+                <div className="mb-2 flex items-center gap-2">
+                  <AlertCircle className="shrink-0" size={16} />
+                  <h6 className="text-sm font-medium">Error Saving Code</h6>
+                </div>
+                <p className="whitespace-pre-wrap text-sm">{saveError}</p>
+              </div>
+            )}
+
+            <div className="flex justify-end gap-2">
+              <Button
+                variant="outline"
+                className="mr-2"
+                onClick={handleResetCode}
+              >
+                Reset Code
+              </Button>
+              <Button
+                onClick={handleSaveCode}
+                disabled={isSaving || !updatedCode || !workflowFile}
+              >
+                Save & Restart Server
+                {isSaving && <Loader2 className="ml-2 h-4 w-4 animate-spin" />}
+              </Button>
+            </div>
+          </div>
+        </div>
+      </div>
+    </>
+  );
+}
@@ -20,7 +20,7 @@
    "paths": {
      "@/*": ["./*"]
    },
-    "target": "ES2017"
+    "target": "ES2022"
  },
  "include": ["next-env.d.ts", "**/*.ts", "**/*.tsx", ".next/types/**/*.ts"],
  "exclude": ["node_modules"]
@@ -1,7 +1,7 @@
 {
  "name": "@llamaindex/server",
  "description": "LlamaIndex Server",
-  "version": "0.2.0",
+  "version": "0.2.2",
  "type": "module",
  "main": "./dist/index.cjs",
  "module": "./dist/index.js",
@@ -27,21 +27,25 @@
    "directory": "packages/server"
  },
  "scripts": {
-    "clean": "rm -rf ./dist ./server next/.next next/out",
+    "dev": "bunchee --watch",
+    "clean": "rm -rf ./dist ./server next/.next next/out ./temp",
    "prebuild": "pnpm clean",
    "build": "bunchee",
-    "postbuild": "pnpm copy:next-src && pnpm build:static && pnpm copy:static",
-    "copy:next-src": "cp -r ./next ./server && pnpm build:css && rm -rf ./server/postcss.config.js",
-    "build:css": "postcss server/app/globals.css -o server/app/globals.css",
-    "build:static": "cd ./next && next build",
-    "copy:static": "cp -r ./next/out ./dist/static",
-    "dev": "bunchee --watch"
+    "postbuild": "pnpm prepare:ts-server && pnpm prepare:py-static",
+    "prepare:ts-server": "pnpm copy:next-src && pnpm build:css && pnpm build:api",
+    "prepare:py-static": "pnpm prepare:static && pnpm build:static && pnpm copy:static",
+    "copy:next-src": "cp -r ./next ./server",
+    "build:css": "postcss server/app/globals.css -o server/app/globals.css && rm -rf ./server/postcss.config.js",
+    "build:api": "rm -rf ./server/app/api && tsc --skipLibCheck --project tsconfig.api.json",
+    "prepare:static": "cp -r ./next ./temp && rm -rf ./temp/app/api && mv ./temp/next-build.config.ts ./temp/next.config.ts",
+    "build:static": "cd ./temp && next build",
+    "copy:static": "cp -r ./temp/out ./dist/static && rm -rf ./temp"
  },
  "devDependencies": {
    "@tailwindcss/postcss": "^4",
    "@types/babel__standalone": "^7.1.9",
    "@types/babel__traverse": "^7.20.7",
-    "llamaindex": "0.10.2",
+    "llamaindex": "~0.11.0",
    "postcss": "^8.5.3",
    "postcss-cli": "^11.0.1",
    "tailwindcss": "^4",
@@ -55,7 +59,7 @@
    "@babel/traverse": "^7.27.0",
    "@babel/types": "^7.27.0",
    "@hookform/resolvers": "^5.0.1",
-    "@llamaindex/chat-ui": "0.4.3",
+    "@llamaindex/chat-ui": "0.4.4",
    "@radix-ui/react-accordion": "^1.2.3",
    "@radix-ui/react-alert-dialog": "^1.1.7",
    "@radix-ui/react-aspect-ratio": "^1.1.3",
@@ -103,9 +107,9 @@
    "vaul": "^1.1.2"
  },
  "peerDependencies": {
-    "@llamaindex/env": "^0.1.29",
-    "@llamaindex/workflow": "^1.1.0",
-    "llamaindex": "^0.10.2",
+    "@llamaindex/env": "~0.1.30",
+    "@llamaindex/workflow": "~1.1.3",
+    "llamaindex": "~0.11.0",
    "zod": "^3.24.2",
    "zod-to-json-schema": "^3.23.3"
  },
@@ -155,10 +155,12 @@ export function extractAllArtifacts(messages: Message[]): Artifact[] {
    const artifacts =
      message.annotations
        ?.filter(
-          (annotation) =>
+          (
+            annotation,
+          ): annotation is z.infer<typeof artifactAnnotationSchema> =>
            artifactAnnotationSchema.safeParse(annotation).success,
        )
-        .map((artifact) => artifact as Artifact) ?? [];
+        .map((annotation) => annotation.data as Artifact) ?? [];

    allArtifacts.push(...artifacts);
  }
@@ -1,30 +0,0 @@
-import { getEnv } from "@llamaindex/env";
-import type { IncomingMessage, ServerResponse } from "http";
-import { LLamaCloudFileService } from "llamaindex";
-import { sendJSONResponse } from "../utils/request";
-
-export const getLlamaCloudConfig = async (
-  req: IncomingMessage,
-  res: ServerResponse,
-) => {
-  if (!getEnv("LLAMA_CLOUD_API_KEY")) {
-    return sendJSONResponse(res, 500, {
-      error: "env variable LLAMA_CLOUD_API_KEY is required to use LlamaCloud",
-    });
-  }
-
-  try {
-    const config = {
-      projects: await LLamaCloudFileService.getAllProjectsWithPipelines(),
-      pipeline: {
-        pipeline: getEnv("LLAMA_CLOUD_INDEX_NAME"),
-        project: getEnv("LLAMA_CLOUD_PROJECT_NAME"),
-      },
-    };
-    return sendJSONResponse(res, 200, config);
-  } catch (error) {
-    return sendJSONResponse(res, 500, {
-      error: "Failed to fetch LlamaCloud configuration",
-    });
-  }
-};
@@ -1,23 +0,0 @@
-import fs from "fs";
-import type { IncomingMessage, ServerResponse } from "http";
-import { promisify } from "util";
-import { sendJSONResponse } from "../utils/request";
-
-export const handleServeFiles = async (
-  req: IncomingMessage,
-  res: ServerResponse,
-  pathname: string,
-) => {
-  const filePath = pathname.substring("/api/files/".length);
-  if (!filePath.startsWith("output") && !filePath.startsWith("data")) {
-    return sendJSONResponse(res, 400, { error: "No permission" });
-  }
-  const decodedFilePath = decodeURIComponent(filePath);
-  const fileExists = await promisify(fs.exists)(decodedFilePath);
-  if (fileExists) {
-    const fileStream = fs.createReadStream(decodedFilePath);
-    fileStream.pipe(res);
-  } else {
-    return sendJSONResponse(res, 404, { error: "File not found" });
-  }
-};
@@ -7,10 +7,8 @@ import path from "path";
 import { parse } from "url";
 import { promisify } from "util";
 import { handleChat } from "./handlers/chat";
-import { getLlamaCloudConfig } from "./handlers/cloud";
-import { getComponents } from "./handlers/components";
-import { handleServeFiles } from "./handlers/files";
 import type { LlamaIndexServerOptions } from "./types";
+
 const nextDir = path.join(__dirname, "..", "server");
 const configFile = path.join(__dirname, "..", "server", "public", "config.js");
 const dev = process.env.NODE_ENV !== "production";
@@ -44,6 +42,7 @@ export class LlamaIndexServer {
        ? "/api/chat/config/llamacloud"
        : undefined;
    const componentsApi = this.componentsDir ? "/api/components" : undefined;
+    const devMode = uiConfig?.devMode ?? false;

    // content in javascript format
    const content = `
@@ -52,7 +51,8 @@ export class LlamaIndexServer {
        APP_TITLE: ${JSON.stringify(appTitle)},
        LLAMA_CLOUD_API: ${JSON.stringify(llamaCloudApi)},
        STARTER_QUESTIONS: ${JSON.stringify(starterQuestions)},
-        COMPONENTS_API: ${JSON.stringify(componentsApi)}
+        COMPONENTS_API: ${JSON.stringify(componentsApi)},
+        DEV_MODE: ${JSON.stringify(devMode)}
      }
    `;
    fs.writeFileSync(configFile, content);
@@ -71,33 +71,25 @@ export class LlamaIndexServer {
    const server = createServer((req, res) => {
      const parsedUrl = parse(req.url!, true);
      const pathname = parsedUrl.pathname;
+      const query = parsedUrl.query;

      if (pathname === "/api/chat" && req.method === "POST") {
+        // because of https://github.com/vercel/next.js/discussions/79402 we can't use route.ts here, so we need to call this custom route
+        // when calling `pnpm eject`, the user will get an equivalent route at [path to chat route.ts]
+        // make sure to keep its semantic in sync with handleChat
        return handleChat(req, res, this.workflowFactory);
      }

-      if (pathname?.startsWith("/api/files") && req.method === "GET") {
-        return handleServeFiles(req, res, pathname);
-      }
-
      if (
        this.componentsDir &&
        pathname === "/api/components" &&
        req.method === "GET"
      ) {
-        return getComponents(req, res, this.componentsDir);
-      }
-
-      if (
-        getEnv("LLAMA_CLOUD_API_KEY") &&
-        pathname === "/api/chat/config/llamacloud" &&
-        req.method === "GET"
-      ) {
-        return getLlamaCloudConfig(req, res);
+        query.componentsDir = this.componentsDir;
      }

      const handle = this.app.getRequestHandler();
-      handle(req, res, parsedUrl);
+      handle(req, res, { ...parsedUrl, query });
    });

    server.listen(this.port, () => {
@@ -17,6 +17,7 @@ export type UIConfig = {
  starterQuestions?: string[];
  componentsDir?: string;
  llamaCloudIndexSelector?: boolean;
+  devMode?: boolean;
 };

 export type LlamaIndexServerOptions = NextAppOptions & {
@@ -0,0 +1,10 @@
+{
+  "extends": "./tsconfig.json",
+  "compilerOptions": {
+    "rootDir": "./next/app/api",
+    "outDir": "./server/app/api",
+    "emitDeclarationOnly": false
+  },
+  "include": ["./next/app/api"],
+  "exclude": ["./next/app/api/chat/route.ts"]
+}
@@ -181,14 +181,14 @@ importers:
        specifier: ^5.0.1
        version: 5.0.1(react-hook-form@7.56.1(react@19.1.0))
      '@llamaindex/chat-ui':
-        specifier: 0.4.3
-        version: 0.4.3(@babel/runtime@7.27.0)(@codemirror/autocomplete@6.18.6)(@codemirror/language@6.11.0)(@codemirror/lint@6.8.5)(@codemirror/search@6.5.10)(@codemirror/state@6.5.2)(@codemirror/theme-one-dark@6.1.2)(@codemirror/view@6.36.7)(@types/react-dom@19.1.2(@types/react@19.1.2))(@types/react@19.1.2)(codemirror@6.0.1)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
+        specifier: 0.4.4
+        version: 0.4.4(@babel/runtime@7.27.0)(@codemirror/autocomplete@6.18.6)(@codemirror/language@6.11.0)(@codemirror/lint@6.8.5)(@codemirror/search@6.5.10)(@codemirror/state@6.5.2)(@codemirror/theme-one-dark@6.1.2)(@codemirror/view@6.36.7)(@types/react-dom@19.1.2(@types/react@19.1.2))(@types/react@19.1.2)(codemirror@6.0.1)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
      '@llamaindex/env':
-        specifier: ^0.1.29
-        version: 0.1.29
+        specifier: ~0.1.30
+        version: 0.1.30
      '@llamaindex/workflow':
-        specifier: ^1.1.0
-        version: 1.1.0(@llamaindex/core@0.6.2)(@llamaindex/env@0.1.29)(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(zod@3.24.3)
+        specifier: ~1.1.3
+        version: 1.1.3(@llamaindex/core@0.6.6)(@llamaindex/env@0.1.30)(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3)
      '@radix-ui/react-accordion':
        specifier: ^1.2.3
        version: 1.2.8(@types/react-dom@19.1.2(@types/react@19.1.2))(@types/react@19.1.2)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
@@ -341,8 +341,8 @@ importers:
        specifier: ^7.20.7
        version: 7.20.7
      llamaindex:
-        specifier: 0.10.2
-        version: 0.10.2(tree-sitter@0.22.4)(web-tree-sitter@0.24.7)(zod@3.24.3)
+        specifier: ~0.11.0
+        version: 0.11.1(@llama-flow/core@0.4.1(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3))(tree-sitter@0.22.4)(web-tree-sitter@0.24.7)(zod@3.24.3)
      postcss:
        specifier: ^8.5.3
        version: 8.5.3
@@ -378,7 +378,7 @@ importers:
        version: 0.0.4(openapi-types@12.1.3)
      '@llamaindex/workflow':
        specifier: 1.1.0
-        version: 1.1.0(@llamaindex/core@0.6.2)(@llamaindex/env@0.1.29)(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(zod@3.24.3)
+        version: 1.1.0(@llamaindex/core@0.6.2)(@llamaindex/env@0.1.29)(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3)
      dotenv:
        specifier: ^16.4.7
        version: 16.5.0
@@ -392,6 +392,9 @@ importers:
      '@types/node':
        specifier: ^20.10.3
        version: 20.17.30
+      nodemon:
+        specifier: ^3.1.10
+        version: 3.1.10
      tsx:
        specifier: ^4.7.2
        version: 4.19.3
@@ -1177,8 +1180,8 @@ packages:
      zod:
        optional: true

-  '@llamaindex/chat-ui@0.4.3':
-    resolution: {integrity: sha512-JeS5pzEXZRRzTL70lHRtbWbaPjeZuMMHlqkTNBgYqc2eMM+o9VMOzVKKsXoYfOeE+wunqJd+cBrtkECGEmhRZg==}
+  '@llamaindex/chat-ui@0.4.4':
+    resolution: {integrity: sha512-sE3mJxlmAV3eiIaqOnioUNYYBLJJL8sy2mdFP4xXDf3PfkcEn0wT2Om/ButpVgIXHlPG9lONtv8mzw7hWq2Atg==}
    peerDependencies:
      react: ^18.2.0 || ^19.0.0 || ^19.0.0-rc

@@ -1188,12 +1191,22 @@ packages:
      '@llamaindex/core': 0.6.2
      '@llamaindex/env': 0.1.29

+  '@llamaindex/cloud@4.0.9':
+    resolution: {integrity: sha512-+tIPghKAzXFvKEPmbJZpcFJgzmxLq/HfSKlJAIinRdrSDBC489oAJsqkt6StxmlFUhWS3ATOtTWpqWZIQ9HtzA==}
+    peerDependencies:
+      '@llama-flow/core': ^0.4.1
+      '@llamaindex/core': 0.6.6
+      '@llamaindex/env': 0.1.30
+
  '@llamaindex/core@0.6.1':
    resolution: {integrity: sha512-OsY8JV2nbfWYjp24HEKDbtM6vOa0TNUreMjDMxUkw5Wet1TxKxjVq0YHY01RoqMbbTJ+Jsxa42FhzTaDV2gMkg==}

  '@llamaindex/core@0.6.2':
    resolution: {integrity: sha512-e0NH7X1C8yq9sCqj1Ys9jL9rITCXBKfZwmKnTHYxpXorPeuNmlFmuoogahJTwLNsILBDTOVAUexO88SqmAFUrA==}

+  '@llamaindex/core@0.6.6':
+    resolution: {integrity: sha512-UplAAxsv3LnU1EP2kPuyqDtRQmQqYtsI00rVDeexOwgqwNTfGddI0yPRrFTBSNe0q2ttBuPsB7D3M7j4UY3yXQ==}
+
  '@llamaindex/env@0.1.29':
    resolution: {integrity: sha512-GoITt+QLDNIhu2i1sGsPH8tHH13Gxp4i4ofMFlefrHagytvF761MG7nvTAQVIqP9kxBYk4dnSQ3lQnVPFAfMZQ==}
    peerDependencies:
@@ -1205,6 +1218,17 @@ packages:
      gpt-tokenizer:
        optional: true

+  '@llamaindex/env@0.1.30':
+    resolution: {integrity: sha512-y6kutMcCevzbmexUgz+HXf7KiZemzAoFEYSjAILfR+cG6FmYSF8XvLbGOB34Kx8mlRi7EI8rZXpezJ5qCqOyZg==}
+    peerDependencies:
+      '@huggingface/transformers': ^3.5.0
+      gpt-tokenizer: ^2.5.0
+    peerDependenciesMeta:
+      '@huggingface/transformers':
+        optional: true
+      gpt-tokenizer:
+        optional: true
+
  '@llamaindex/node-parser@2.0.2':
    resolution: {integrity: sha512-pWqEfFRMW5TFzbDyEDWqhipR9P/HBF2C+pEeTroyRACZtKJLbeg3R3e2g4gmfnPxmxQIXkGn2UFK8Io4nwVNOg==}
    peerDependencies:
@@ -1213,6 +1237,14 @@ packages:
      tree-sitter: ^0.22.0
      web-tree-sitter: ^0.24.3

+  '@llamaindex/node-parser@2.0.6':
+    resolution: {integrity: sha512-8vJ7hsh1FKfAfO90nNe/G+/E2hM31dT/+1+0cPpZp02Ez98Jhwj+vTcCRlvyZznCn6DzHNBuw+uP1wKCqylSnQ==}
+    peerDependencies:
+      '@llamaindex/core': 0.6.6
+      '@llamaindex/env': 0.1.30
+      tree-sitter: ^0.22.0
+      web-tree-sitter: ^0.24.3
+
  '@llamaindex/openai@0.2.1':
    resolution: {integrity: sha512-gSBYL1vMaEihXm5K5gIHpUmvDS78ACPgQPlwf/ZnE82UD/vFNOWOqQbwESJKcygC0lHrwhrKJAuBAzXSNcpznA==}

@@ -1252,6 +1284,13 @@ packages:
      '@llamaindex/env': 0.1.29
      zod: ^3.23.8

+  '@llamaindex/workflow@1.1.3':
+    resolution: {integrity: sha512-B9wlN4xMtfbnIQW84wTC0GdyIaLGXYnZ4I0LENV2UM/0wqlXKI7YoBTO3cJvi6pO1sPA+K29Tqw822VsGF+GkA==}
+    peerDependencies:
+      '@llamaindex/core': 0.6.6
+      '@llamaindex/env': 0.1.30
+      zod: ^3.23.8
+
  '@manypkg/find-root@1.1.0':
    resolution: {integrity: sha512-mki5uBvhHzO8kYYix/WRy2WX8S3B5wdVSc9D6KcU5lQNglP2yt58/VfLuAK49glRXChosY8ap2oJ1qgma3GUVA==}

@@ -2481,6 +2520,9 @@ packages:
  '@types/responselike@1.0.3':
    resolution: {integrity: sha512-H/+L+UkTV33uf49PH5pCAUBVPNj2nDBXTN+qS1dOwyyg24l3CcicicCA7ca+HMvJBZcFgl5r8e+RR6elsb4Lyw==}

+  '@types/retry@0.12.2':
+    resolution: {integrity: sha512-XISRgDJ2Tc5q4TRqvgJtzsRkFYNJzZrhTdtMoGVBttwzzQJkPnS3WWTFc7kuDRoPtPakl+T+OfdEUjYJj7Jbow==}
+
  '@types/retry@0.12.5':
    resolution: {integrity: sha512-3xSjTp3v03X/lSQLkczaN9UIEwJMoMCA1+Nb5HfbJEQWogdeQIyVtTvxPXDQjZ5zws8rFQfVfRdz03ARihPJgw==}

@@ -3882,6 +3924,10 @@ packages:
    resolution: {integrity: sha512-R3pbpkcIqv2Pm3dUwgjclDRVmWpTJW2DcMzcIhEXEx1oh/CEMObMm3KLmRJOdvhM7o4uQBnwr8pzRK2sJWIqfg==}
    engines: {node: '>= 0.4'}

+  has-flag@3.0.0:
+    resolution: {integrity: sha512-sKJf1+ceQBr4SMkvQnBDNDtf4TXpVhVGateu0t918bl30FnbE2m4vNLX+VWe/dpjlb+HugGYzW7uQXH98HPEYw==}
+    engines: {node: '>=4'}
+
  has-flag@4.0.0:
    resolution: {integrity: sha512-EykJT/Q1KjTWctppgIAgfSO0tKVuZUjhgMr17kqTumMl6Afv3EISleU7qZUzoXDFTAHTDC4NOoG/ZxU3EvlMPQ==}
    engines: {node: '>=8'}
@@ -3987,6 +4033,9 @@ packages:
  ieee754@1.2.1:
    resolution: {integrity: sha512-dcyqhDvX1C46lXZcVqCpK+FtMRQVdIMN6/Df5js2zouUsqG7I6sFxitIC+7KYK29KdXOLHdu9zL4sFnoVQnqaA==}

+  ignore-by-default@1.0.1:
+    resolution: {integrity: sha512-Ius2VYcGNk7T90CppJqcIkS5ooHUZyIQK+ClZfMfMNFEF9VSE73Fq+906u/CWu92x4gzZMWOwfFYckPObzdEbA==}
+
  ignore@5.3.2:
    resolution: {integrity: sha512-hsBTNUqQTDwkWtcdYI2i06Y/nUBEsNEDJKjWdigLvegy8kDuJAS8uRlpkkcQpyEXL0Z/pjDy5HBmMjRCJ2gq+g==}
    engines: {node: '>= 4'}
@@ -4117,6 +4166,10 @@ packages:
  is-module@1.0.0:
    resolution: {integrity: sha512-51ypPSPCoTEIN9dy5Oy+h4pShgJmPCygKfyRCISBI+JoWT/2oJvK8QPxmwv7b/p239jXrm9M1mlQbyKJ5A152g==}

+  is-network-error@1.1.0:
+    resolution: {integrity: sha512-tUdRRAnhT+OtCZR/LxZelH/C7QtjtFrTu5tXCA8pl55eTUElUHT+GPYV8MBMBvea/j+NxQqVt3LbWMRir7Gx9g==}
+    engines: {node: '>=16'}
+
  is-number-object@1.1.1:
    resolution: {integrity: sha512-lZhclumE1G6VYD8VHe35wFaIif+CTy5SJIi5+3y4psDgWu4wPDoBhF8NxUOinEc7pHgiTsT6MaBb92rKhhD+Xw==}
    engines: {node: '>= 0.4'}
@@ -4387,6 +4440,10 @@ packages:
    resolution: {integrity: sha512-tUbb+HYaQ+WP1f4NrYnEOIgPPRUePh3sPnOf/wbxTYAyZTnREpSiwCfLUGgYOiIYKK5PS7Oo/eTZCP7jpi7gSA==}
    engines: {node: '>=18.0.0'}

+  llamaindex@0.11.1:
+    resolution: {integrity: sha512-d+sMHc6b/Lw8S1HhTVtR5hgEkg4uCpADpgZDcpHa/QlSQXvCmSkEdI9DOhoTDMhNWWatKnQZADPDuVzkyysQ3w==}
+    engines: {node: '>=18.0.0'}
+
  locate-path@5.0.0:
    resolution: {integrity: sha512-t7hw9pI+WvuwNJXwk5zVHpyhIqzg2qTlklJOf0mVxGSbe3Fp2VieZcduNYjaLDoy6p9uGpQEGWG87WpMKlNq8g==}
    engines: {node: '>=8'}
@@ -4805,6 +4862,11 @@ packages:
    resolution: {integrity: sha512-LA4ZjwlnUblHVgq0oBF3Jl/6h/Nvs5fzBLwdEF4nuxnFdsfajde4WfxtJr3CaiH+F6ewcIB/q4jQ4UzPyid+CQ==}
    hasBin: true

+  nodemon@3.1.10:
+    resolution: {integrity: sha512-WDjw3pJ0/0jMFmyNDp3gvY2YizjLmmOUQo6DEBY+JgdvW/yQ9mEeSw6H5ythl5Ny2ytb7f9C2nIbjSxMNzbJXw==}
+    engines: {node: '>=10'}
+    hasBin: true
+
  nopt@5.0.0:
    resolution: {integrity: sha512-Tbj67rffqceeLpcRXrT7vKAN8CwfPeIBgM7E6iBkmKLV7bEMwpGgYLGv0jACUsECaa/vuxP0IjEont6umdMgtQ==}
    engines: {node: '>=6'}
@@ -4962,6 +5024,10 @@ packages:
    resolution: {integrity: sha512-y3b8Kpd8OAN444hxfBbFfj1FY/RjtTd8tzYwhUqNYXx0fXx2iX4maP4Qr6qhIKbQXI02wTLAda4fYUbDagTUFw==}
    engines: {node: '>=6'}

+  p-retry@6.2.1:
+    resolution: {integrity: sha512-hEt02O4hUct5wtwg4H4KcWgDdm+l1bOaEy/hWzd8xtXB9BqxTWBBhb+2ImAtH4Cv4rPjV76xN3Zumqk3k3AhhQ==}
+    engines: {node: '>=16.17'}
+
  p-timeout@3.2.0:
    resolution: {integrity: sha512-rhIwUycgwwKcP9yTOOFK/AKsAopjjCakVqLHePO3CC6Mir1Z99xT+R63jZxAT5lFZLa2inS5h+ZS2GvR99/FBg==}
    engines: {node: '>=8'}
@@ -5258,6 +5324,9 @@ packages:
  proxy-from-env@1.1.0:
    resolution: {integrity: sha512-D+zkORCbA9f1tdWRK0RaCR3GPv50cMxcrz4X8k5LTSUD1Dkw47mKJEZQNunItRTkWwgtaUSo1RVFRIG9ZXiFYg==}

+  pstree.remy@1.1.8:
+    resolution: {integrity: sha512-77DZwxQmxKnu3aR542U+X8FypNzbfJ+C5XQDk3uWjWxn6151aIMGthWYRXTqT1E5oJvg+ljaa2OJi+VfvCOQ8w==}
+
  pump@3.0.2:
    resolution: {integrity: sha512-tUPXtzlGM8FE3P0ZL6DVs/3P58k9nk8/jZeQCurTJylQA8qFYzHFfhBJkuqyE0FifOsQ0uKWekiZ5g8wtr28cw==}

@@ -5503,6 +5572,10 @@ packages:
    resolution: {integrity: sha512-9LkiTwjUh6rT555DtE9rTX+BKByPfrMzEAtnlEtdEwr3Nkffwiihqe2bWADg+OQRjt9gl6ICdmB/ZFDCGAtSow==}
    engines: {node: '>= 4'}

+  retry@0.13.1:
+    resolution: {integrity: sha512-XQBQ3I8W1Cge0Seh+6gjj03LbmRFWuoszgK9ooCpwYIrhhoO80pfq4cUkU5DkknwfOfFteRwlZ56PYOGYyFWdg==}
+    engines: {node: '>= 4'}
+
  reusify@1.1.0:
    resolution: {integrity: sha512-g6QUff04oZpHs0eG5p83rFLhHeV00ug/Yf9nZM6fLeUrPguBTkTQOdpAWWspMh55TZfVQDPaN3NQJfbVRAxdIw==}
    engines: {iojs: '>=1.0.0', node: '>=0.10.0'}
@@ -5674,6 +5747,10 @@ packages:
  simple-swizzle@0.2.2:
    resolution: {integrity: sha512-JA//kQgZtbuY83m+xT+tXJkmJncGMTFT+C+g2h2R9uxkYIrE2yy9sgmcLhCnw57/WSD+Eh3J97FPEDFnbXnDUg==}

+  simple-update-notifier@2.0.0:
+    resolution: {integrity: sha512-a2B9Y0KlNXl9u/vsW6sTIu9vGEpfKu2wRV6l1H3XEas/0gUIzGzBoP/IouTcUQbm9JWZLH3COxyn03TYlFax6w==}
+    engines: {node: '>=10'}
+
  sisteransi@1.0.5:
    resolution: {integrity: sha512-bLGGlR1QxBcynn2d5YmDX4MGjlZvy2MRBDRNHLJ8VI6l6+9FUiyTFNJ0IveOSP0bcXgVDPRcfGqA0pjaqUpfVg==}

@@ -5832,6 +5909,10 @@ packages:
      babel-plugin-macros:
        optional: true

+  supports-color@5.5.0:
+    resolution: {integrity: sha512-QjVjwdXIt408MIiAqCX4oUKsgU2EqAGzs2Ppkm4aQYbjm+ZEWEcW4SfFNTr4uMNZma0ey4f5lgLrkB0aX0QMow==}
+    engines: {node: '>=4'}
+
  supports-color@7.2.0:
    resolution: {integrity: sha512-qpCAvRl9stuOHveKsn7HncJRvv501qIacKzQlO/+Lwxc9+0q2wLyv4Dfvt80/DPn2pqOBsJdDiogXGR9+OvwRw==}
    engines: {node: '>=8'}
@@ -5928,6 +6009,10 @@ packages:
  to-vfile@6.1.0:
    resolution: {integrity: sha512-BxX8EkCxOAZe+D/ToHdDsJcVI4HqQfmw0tCkp31zf3dNP/XWIAjU4CmeuSwsSoOzOTqHPOL0KUzyZqJplkD0Qw==}

+  touch@3.1.1:
+    resolution: {integrity: sha512-r0eojU4bI8MnHr8c5bNo7lJDdI2qXlWWJk6a9EAFG7vbhTjElYhBVS3/miuE0uOuoLdb8Mc/rVfsmm6eo5o9GA==}
+    hasBin: true
+
  tr46@0.0.3:
    resolution: {integrity: sha512-N3WMsuqV66lT30CrXNbEjx4GEwlow3v6rr4mCcv6prnfwhS01rkgyFdjPNBYd9br7LpXV1+Emh01fHnq2Gdgrw==}

@@ -6018,6 +6103,9 @@ packages:
    resolution: {integrity: sha512-nWJ91DjeOkej/TA8pXQ3myruKpKEYgqvpw9lz4OPHj/NWFNluYrjbz9j01CJ8yKQd2g4jFoOkINCTW2I5LEEyw==}
    engines: {node: '>= 0.4'}

+  undefsafe@2.0.5:
+    resolution: {integrity: sha512-WxONCrssBM8TSPRqN5EmsjVrsv4A8X12J4ArBiiayv3DyyG3ZlIg6yysuuSYdZsVz3TKcTg2fd//Ujd4CHV1iA==}
+
  underscore@1.13.7:
    resolution: {integrity: sha512-GMXzWtsc57XAtguZgaQViUOzs0KTkk8ojr3/xAxXLITqf/3EMwxC0inyETfDFjH/Krbhuep0HNbbjI9i/q3F3g==}

@@ -6348,6 +6436,9 @@ packages:
  zod@3.24.3:
    resolution: {integrity: sha512-HhY1oqzWCQWuUqvBFnsyrtZRhyPeR7SUGv+C4+MsisMuVfSPx8HpwWqH8tRahSlt6M3PiFAcoeFhZAqIXTxoSg==}

+  zod@3.25.13:
+    resolution: {integrity: sha512-Q8mvk2iWi7rTDfpQBsu4ziE7A6AxgzJ5hzRyRYQkoV3A3niYsXVwDaP1Kbz3nWav6S+VZ6k2OznFn8ZyDHvIrg==}
+
  zwitch@2.0.4:
    resolution: {integrity: sha512-bXE4cR/kVZhKZX/RjPEflHaKVhUVl85noU3v6b8apfQEc1x4A+zBxjZ4lN8LqGd6WZ3dl98pY4o717VFmoPp+A==}

@@ -6462,7 +6553,7 @@ snapshots:
      '@babel/parser': 7.27.0
      '@babel/template': 7.27.0
      '@babel/types': 7.27.0
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      globals: 11.12.0
    transitivePeerDependencies:
      - supports-color
@@ -6891,7 +6982,7 @@ snapshots:
  '@eslint/config-array@0.19.2':
    dependencies:
      '@eslint/object-schema': 2.1.6
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      minimatch: 3.1.2
    transitivePeerDependencies:
      - supports-color
@@ -6909,7 +7000,7 @@ snapshots:
  '@eslint/eslintrc@3.3.1':
    dependencies:
      ajv: 6.12.6
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      espree: 10.3.0
      globals: 14.0.0
      ignore: 5.3.2
@@ -7108,12 +7199,13 @@ snapshots:
      '@lezer/highlight': 1.2.1
      '@lezer/lr': 1.4.2

-  '@llama-flow/core@0.4.1(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(zod@3.24.3)':
+  '@llama-flow/core@0.4.1(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3)':
    optionalDependencies:
      next: 15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)
+      p-retry: 6.2.1
      zod: 3.24.3

-  '@llamaindex/chat-ui@0.4.3(@babel/runtime@7.27.0)(@codemirror/autocomplete@6.18.6)(@codemirror/language@6.11.0)(@codemirror/lint@6.8.5)(@codemirror/search@6.5.10)(@codemirror/state@6.5.2)(@codemirror/theme-one-dark@6.1.2)(@codemirror/view@6.36.7)(@types/react-dom@19.1.2(@types/react@19.1.2))(@types/react@19.1.2)(codemirror@6.0.1)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)':
+  '@llamaindex/chat-ui@0.4.4(@babel/runtime@7.27.0)(@codemirror/autocomplete@6.18.6)(@codemirror/language@6.11.0)(@codemirror/lint@6.8.5)(@codemirror/search@6.5.10)(@codemirror/state@6.5.2)(@codemirror/theme-one-dark@6.1.2)(@codemirror/view@6.36.7)(@types/react-dom@19.1.2(@types/react@19.1.2))(@types/react@19.1.2)(codemirror@6.0.1)(react-dom@19.1.0(react@19.1.0))(react@19.1.0)':
    dependencies:
      '@codemirror/lang-css': 6.3.1
      '@codemirror/lang-html': 6.4.9
@@ -7173,6 +7265,14 @@ snapshots:
      '@llamaindex/core': 0.6.2
      '@llamaindex/env': 0.1.29

+  '@llamaindex/cloud@4.0.9(@llama-flow/core@0.4.1(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3))(@llamaindex/core@0.6.6)(@llamaindex/env@0.1.30)':
+    dependencies:
+      '@llama-flow/core': 0.4.1(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3)
+      '@llamaindex/core': 0.6.6
+      '@llamaindex/env': 0.1.30
+      p-retry: 6.2.1
+      zod: 3.25.13
+
  '@llamaindex/core@0.6.1':
    dependencies:
      '@llamaindex/env': 0.1.29
@@ -7195,12 +7295,29 @@ snapshots:
      - '@huggingface/transformers'
      - gpt-tokenizer

+  '@llamaindex/core@0.6.6':
+    dependencies:
+      '@llamaindex/env': 0.1.30
+      '@types/node': 22.15.2
+      magic-bytes.js: 1.12.1
+      zod: 3.24.3
+      zod-to-json-schema: 3.24.5(zod@3.24.3)
+    transitivePeerDependencies:
+      - '@huggingface/transformers'
+      - gpt-tokenizer
+
  '@llamaindex/env@0.1.29':
    dependencies:
      '@aws-crypto/sha256-js': 5.2.0
      js-tiktoken: 1.0.20
      pathe: 1.1.2

+  '@llamaindex/env@0.1.30':
+    dependencies:
+      '@aws-crypto/sha256-js': 5.2.0
+      js-tiktoken: 1.0.20
+      pathe: 1.1.2
+
  '@llamaindex/node-parser@2.0.2(@llamaindex/core@0.6.2)(@llamaindex/env@0.1.29)(tree-sitter@0.22.4)(web-tree-sitter@0.24.7)':
    dependencies:
      '@llamaindex/core': 0.6.2
@@ -7209,6 +7326,14 @@ snapshots:
      tree-sitter: 0.22.4
      web-tree-sitter: 0.24.7

+  '@llamaindex/node-parser@2.0.6(@llamaindex/core@0.6.6)(@llamaindex/env@0.1.30)(tree-sitter@0.22.4)(web-tree-sitter@0.24.7)':
+    dependencies:
+      '@llamaindex/core': 0.6.6
+      '@llamaindex/env': 0.1.30
+      html-to-text: 9.0.5
+      tree-sitter: 0.22.4
+      web-tree-sitter: 0.24.7
+
  '@llamaindex/openai@0.2.1':
    dependencies:
      '@llamaindex/core': 0.6.1
@@ -7286,9 +7411,15 @@ snapshots:
      '@llamaindex/env': 0.1.29
      zod: 3.24.3

-  '@llamaindex/workflow@1.1.0(@llamaindex/core@0.6.2)(@llamaindex/env@0.1.29)(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(zod@3.24.3)':
+  '@llamaindex/workflow@1.0.3(@llamaindex/core@0.6.6)(@llamaindex/env@0.1.30)(zod@3.24.3)':
    dependencies:
-      '@llama-flow/core': 0.4.1(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(zod@3.24.3)
+      '@llamaindex/core': 0.6.6
+      '@llamaindex/env': 0.1.30
+      zod: 3.24.3
+
+  '@llamaindex/workflow@1.1.0(@llamaindex/core@0.6.2)(@llamaindex/env@0.1.29)(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3)':
+    dependencies:
+      '@llama-flow/core': 0.4.1(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3)
      '@llamaindex/core': 0.6.2
      '@llamaindex/env': 0.1.29
      zod: 3.24.3
@@ -7299,6 +7430,19 @@ snapshots:
      - p-retry
      - rxjs

+  '@llamaindex/workflow@1.1.3(@llamaindex/core@0.6.6)(@llamaindex/env@0.1.30)(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3)':
+    dependencies:
+      '@llama-flow/core': 0.4.1(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3)
+      '@llamaindex/core': 0.6.6
+      '@llamaindex/env': 0.1.30
+      zod: 3.24.3
+    transitivePeerDependencies:
+      - '@modelcontextprotocol/sdk'
+      - hono
+      - next
+      - p-retry
+      - rxjs
+
  '@manypkg/find-root@1.1.0':
    dependencies:
      '@babel/runtime': 7.27.0
@@ -8495,6 +8639,8 @@ snapshots:
    dependencies:
      '@types/node': 20.17.30

+  '@types/retry@0.12.2': {}
+
  '@types/retry@0.12.5': {}

  '@types/tar@6.1.5':
@@ -8531,7 +8677,7 @@ snapshots:
      '@typescript-eslint/types': 8.31.0
      '@typescript-eslint/typescript-estree': 8.31.0(typescript@5.8.3)
      '@typescript-eslint/visitor-keys': 8.31.0
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      eslint: 9.22.0(jiti@2.4.2)
      typescript: 5.8.3
    transitivePeerDependencies:
@@ -8546,7 +8692,7 @@ snapshots:
    dependencies:
      '@typescript-eslint/typescript-estree': 8.31.0(typescript@5.8.3)
      '@typescript-eslint/utils': 8.31.0(eslint@9.22.0(jiti@2.4.2))(typescript@5.8.3)
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      eslint: 9.22.0(jiti@2.4.2)
      ts-api-utils: 2.1.0(typescript@5.8.3)
      typescript: 5.8.3
@@ -8559,7 +8705,7 @@ snapshots:
    dependencies:
      '@typescript-eslint/types': 8.31.0
      '@typescript-eslint/visitor-keys': 8.31.0
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      fast-glob: 3.3.3
      is-glob: 4.0.3
      minimatch: 9.0.5
@@ -8743,7 +8889,7 @@ snapshots:

  agent-base@6.0.2:
    dependencies:
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
    transitivePeerDependencies:
      - supports-color
    optional: true
@@ -9101,7 +9247,7 @@ snapshots:
      normalize-path: 3.0.0
      readdirp: 3.6.0
    optionalDependencies:
-      fsevents: 2.3.2
+      fsevents: 2.3.3

  chownr@1.1.4:
    optional: true
@@ -9297,9 +9443,11 @@ snapshots:
    dependencies:
      ms: 2.1.3

-  debug@4.4.0:
+  debug@4.4.0(supports-color@5.5.0):
    dependencies:
      ms: 2.1.3
+    optionalDependencies:
+      supports-color: 5.5.0

  decimal.js-light@2.5.1: {}

@@ -9665,7 +9813,7 @@ snapshots:
  eslint-import-resolver-typescript@3.10.1(eslint-plugin-import@2.31.0)(eslint@9.22.0(jiti@2.4.2)):
    dependencies:
      '@nolyfill/is-core-module': 1.0.39
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      eslint: 9.22.0(jiti@2.4.2)
      get-tsconfig: 4.10.0
      is-bun-module: 2.0.0
@@ -9789,7 +9937,7 @@ snapshots:
      ajv: 6.12.6
      chalk: 4.1.2
      cross-spawn: 7.0.6
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      escape-string-regexp: 4.0.0
      eslint-scope: 8.3.0
      eslint-visitor-keys: 4.2.0
@@ -10168,6 +10316,8 @@ snapshots:

  has-bigints@1.1.0: {}

+  has-flag@3.0.0: {}
+
  has-flag@4.0.0: {}

  has-property-descriptors@1.0.2:
@@ -10278,7 +10428,7 @@ snapshots:
  https-proxy-agent@5.0.1:
    dependencies:
      agent-base: 6.0.2
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
    transitivePeerDependencies:
      - supports-color
    optional: true
@@ -10304,6 +10454,8 @@ snapshots:
  ieee754@1.2.1:
    optional: true

+  ignore-by-default@1.0.1: {}
+
  ignore@5.3.2: {}

  immediate@3.0.6: {}
@@ -10428,6 +10580,8 @@ snapshots:

  is-module@1.0.0: {}

+  is-network-error@1.1.0: {}
+
  is-number-object@1.1.1:
    dependencies:
      call-bound: 1.0.4
@@ -10664,7 +10818,7 @@ snapshots:
    dependencies:
      chalk: 5.4.1
      commander: 13.1.0
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      execa: 8.0.1
      lilconfig: 3.1.3
      listr2: 8.3.2
@@ -10706,6 +10860,26 @@ snapshots:
      - ws
      - zod

+  llamaindex@0.11.1(@llama-flow/core@0.4.1(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3))(tree-sitter@0.22.4)(web-tree-sitter@0.24.7)(zod@3.24.3):
+    dependencies:
+      '@llamaindex/cloud': 4.0.9(@llama-flow/core@0.4.1(next@15.3.1(@opentelemetry/api@1.9.0)(@playwright/test@1.52.0)(react-dom@19.1.0(react@19.1.0))(react@19.1.0))(p-retry@6.2.1)(zod@3.24.3))(@llamaindex/core@0.6.6)(@llamaindex/env@0.1.30)
+      '@llamaindex/core': 0.6.6
+      '@llamaindex/env': 0.1.30
+      '@llamaindex/node-parser': 2.0.6(@llamaindex/core@0.6.6)(@llamaindex/env@0.1.30)(tree-sitter@0.22.4)(web-tree-sitter@0.24.7)
+      '@llamaindex/workflow': 1.0.3(@llamaindex/core@0.6.6)(@llamaindex/env@0.1.30)(zod@3.24.3)
+      '@types/lodash': 4.17.16
+      '@types/node': 22.15.2
+      ajv: 8.17.1
+      lodash: 4.17.21
+      magic-bytes.js: 1.12.1
+    transitivePeerDependencies:
+      - '@huggingface/transformers'
+      - '@llama-flow/core'
+      - gpt-tokenizer
+      - tree-sitter
+      - web-tree-sitter
+      - zod
+
  locate-path@5.0.0:
    dependencies:
      p-locate: 4.1.0
@@ -11115,7 +11289,7 @@ snapshots:
  micromark@3.2.0:
    dependencies:
      '@types/debug': 4.1.12
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      decode-named-character-reference: 1.1.0
      micromark-core-commonmark: 1.1.0
      micromark-factory-space: 1.1.0
@@ -11260,6 +11434,19 @@ snapshots:

  node-gyp-build@4.8.4: {}

+  nodemon@3.1.10:
+    dependencies:
+      chokidar: 3.6.0
+      debug: 4.4.0(supports-color@5.5.0)
+      ignore-by-default: 1.0.1
+      minimatch: 3.1.2
+      pstree.remy: 1.1.8
+      semver: 7.7.1
+      simple-update-notifier: 2.0.0
+      supports-color: 5.5.0
+      touch: 3.1.1
+      undefsafe: 2.0.5
+
  nopt@5.0.0:
    dependencies:
      abbrev: 1.1.1
@@ -11430,6 +11617,12 @@ snapshots:

  p-map@2.1.0: {}

+  p-retry@6.2.1:
+    dependencies:
+      '@types/retry': 0.12.2
+      is-network-error: 1.1.0
+      retry: 0.13.1
+
  p-timeout@3.2.0:
    dependencies:
      p-finally: 1.0.0
@@ -11672,6 +11865,8 @@ snapshots:

  proxy-from-env@1.1.0: {}

+  pstree.remy@1.1.8: {}
+
  pump@3.0.2:
    dependencies:
      end-of-stream: 1.4.4
@@ -11992,6 +12187,8 @@ snapshots:

  retry@0.12.0: {}

+  retry@0.13.1: {}
+
  reusify@1.1.0: {}

  rfdc@1.4.1: {}
@@ -12235,6 +12432,10 @@ snapshots:
      is-arrayish: 0.3.2
    optional: true

+  simple-update-notifier@2.0.0:
+    dependencies:
+      semver: 7.7.1
+
  sisteransi@1.0.5: {}

  slash@3.0.0: {}
@@ -12393,6 +12594,10 @@ snapshots:
      client-only: 0.0.1
      react: 19.1.0

+  supports-color@5.5.0:
+    dependencies:
+      has-flag: 3.0.0
+
  supports-color@7.2.0:
    dependencies:
      has-flag: 4.0.0
@@ -12489,6 +12694,8 @@ snapshots:
      is-buffer: 2.0.5
      vfile: 4.2.1

+  touch@3.1.1: {}
+
  tr46@0.0.3: {}

  tree-sitter@0.22.4:
@@ -12593,6 +12800,8 @@ snapshots:
      has-symbols: 1.1.0
      which-boxed-primitive: 1.1.1

+  undefsafe@2.0.5: {}
+
  underscore@1.13.7: {}

  undici-types@5.26.5: {}
@@ -12820,7 +13029,7 @@ snapshots:
  vite-node@2.1.9(@types/node@20.17.30)(lightningcss@1.29.2):
    dependencies:
      cac: 6.7.14
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      es-module-lexer: 1.7.0
      pathe: 1.1.2
      vite: 5.4.18(@types/node@20.17.30)(lightningcss@1.29.2)
@@ -12855,7 +13064,7 @@ snapshots:
      '@vitest/spy': 2.1.9
      '@vitest/utils': 2.1.9
      chai: 5.2.0
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
      expect-type: 1.2.1
      magic-string: 0.30.17
      pathe: 1.1.2
@@ -12886,7 +13095,7 @@ snapshots:
    dependencies:
      chalk: 4.1.2
      commander: 9.5.0
-      debug: 4.4.0
+      debug: 4.4.0(supports-color@5.5.0)
    transitivePeerDependencies:
      - supports-color

@@ -13023,4 +13232,6 @@ snapshots:

  zod@3.24.3: {}

+  zod@3.25.13: {}
+
  zwitch@2.0.4: {}
@@ -72,7 +72,7 @@ app = LlamaIndexServer(

 The LlamaIndexServer accepts the following configuration parameters:

- `workflow_factory`: A callable that creates a workflow instance for each request
+- `workflow_factory`: A callable that creates a workflow instance for each request. See [Workflow factory contract](#workflow-factory-contract) for more details.
 - `logger`: Optional logger instance (defaults to uvicorn logger)
 - `use_default_routers`: Whether to include default routers (chat, static file serving)
 - `env`: Environment setting ('dev' enables CORS and UI by default)
@@ -83,10 +83,36 @@ The LlamaIndexServer accepts the following configuration parameters:
  - `ui_path`: Path for downloaded UI static files (default: ".ui")
  - `component_dir`: The directory for custom UI components rendering events emitted by the workflow. The default is None, which does not render custom UI components.
  - `llamacloud_index_selector`: Whether to show the LlamaCloud index selector in the chat UI (default: False). Requires `LLAMA_CLOUD_API_KEY` to be set.
+  - `dev_mode`: When enabled, you can update workflow code in the UI and see the changes immediately. It's currently in beta and only supports updating workflow code at `app/workflow.py`. You might also need to set `env="dev"` and start the server with the reload feature enabled.
 - `verbose`: Enable verbose logging
 - `api_prefix`: API route prefix (default: "/api")
 - `server_url`: The deployment URL of the server (default is None)

+## Workflow factory contract
+
+The `workflow_factory` provided will be called for each chat request to initialize a new workflow instance. Additionally, we provide the [ChatRequest](https://github.com/run-llama/create-llama/blob/afe9e9fc16427d20e1dfb635a45e7ed4b46285cb/python/llama-index-server/llama_index/server/api/models.py#L32) object, which includes the request information that is helpful for initializing the workflow. For example:
+```python
+def create_workflow(chat_request: ChatRequest) -> Workflow:
+    # using messages from the chat request to initialize the workflow
+    return MyCustomWorkflow(chat_request.messages)
+```
+
+Your workflow will be executed once for each chat request with the following input parameters are included in workflow's `StartEvent`:
+- `user_msg` [str]: The current user message
+- `chat_history` [list[[ChatMessage](https://docs.llamaindex.ai/en/stable/api_reference/prompts/#llama_index.core.prompts.ChatMessage)]]: All the previous messages of the conversation
+
+Example:
+```python
+@step
+def handle_start_event(ev: StartEvent) -> MyNextEvent:
+    user_msg = ev.user_msg
+    chat_history = ev.chat_history
+    ...
+```
+
+Your workflows can emit `UIEvent` events to render [Custom UI Components](https://github.com/run-llama/create-llama/blob/main/python/llama-index-server/docs/custom_ui_component.md) in the chat UI to improve the user experience.
+Furthermore, you can send `ArtifactEvent` events to render code or document [Artifacts](https://github.com/run-llama/create-llama/blob/main/python/llama-index-server/docs/custom_artifact_event.md) in a dedicated Canvas panel in the chat UI.
+
 ## Default Routers and Features

 ### Chat Router
@@ -107,11 +133,6 @@ When enabled, the server provides a chat interface at the root path (`/`) with:
 - Real-time chat interface
 - API endpoint integration

-### Custom UI Components
-
-You can add custom UI components for your workflow by providing `component_dir` config and adding custom .jsx or .tsx files to the directory.
-See [Custom UI Components](https://github.com/run-llama/create-llama/blob/main/llama-index-server/docs/custom_ui_component.md) for more details.
-
 ## Development Mode

 In development mode (`env="dev"`), the server:
@@ -120,6 +141,20 @@ In development mode (`env="dev"`), the server:
 - Automatically includes the chat UI
 - Provides more verbose logging

+### Workflow Editor (Beta)
+
+In development mode, you can set `dev_mode` to `True` in the UI configuration to enable the workflow editor, which allows you to edit the workflow code directly in the browser.
+
+```python
+app = LlamaIndexServer(
+    workflow_factory=create_workflow,
+    env="dev",
+    ui_config={"dev_mode": True},
+)
+```
+
+**Note**: The workflow editor is currently in beta and only supports updating LlamaIndexServer projects created with [create-llama](https://github.com/run-llama/create-llama/). You also need to start the server via `fastapi dev` so that the server can hot reload the workflow code.
+
 ## API Endpoints

 The server provides the following default endpoints:
@@ -130,11 +165,10 @@ The server provides the following default endpoints:

 ## Best Practices

-1. Always provide a workflow factory that creates fresh workflow instances
-2. Use environment variables for sensitive configuration
-3. Enable verbose logging during development
-4. Configure CORS appropriately for your deployment environment
-5. Use starter questions to guide users in the chat UI
+1. Use environment variables for sensitive configuration
+2. Enable verbose logging during development
+3. Configure CORS appropriately for your deployment environment
+4. Use starter questions to guide users in the chat UI

 ## Getting Started with a New Project

@@ -0,0 +1,59 @@
+# Sending Artifacts to the UI
+
+In addition to UI events for custom components, LlamaIndex Server supports a special `ArtifactEvent` to send structured data like generated documents or code snippets to the UI. These artifacts are displayed in a dedicated "Canvas" panel in the chat interface.
+
+## Artifact Event Structure
+
+To send an artifact, your workflow needs to emit an event with `type: "artifact"`. The `data` payload of this event should include:
+
+- `type`: An `ArtifactType` enum indicating the type of artifact (e.g., `ArtifactType.DOCUMENT`, `ArtifactType.CODE`).
+- `created_at`: A timestamp (e.g., `int(time.time())`) indicating when the artifact was created.
+- `data`: An object containing the specific details of the artifact. The structure of this object depends on the artifact `type`. For example, `DocumentArtifactData` or `CodeArtifactData`.
+
+## Defining and Sending an ArtifactEvent
+
+First, import the necessary classes:
+
+```python
+import time
+from llama_index.server.api.models import (
+    Artifact,
+    ArtifactEvent,
+    ArtifactType,
+    DocumentArtifactData,
+    # CodeArtifactData, # Import if sending code artifacts
+)
+```
+
+Then, within your workflow logic, use `ctx.write_event_to_stream` to emit the event. Here's an example of sending a document artifact, taken from [document_workflow.py](/python/llama-index-server/examples/artifact/document_workflow.py):
+
+```python
+# Assuming 'ctx' is the workflow Context and 'content' is a markdown string
+
+ctx.write_event_to_stream(
+    ArtifactEvent(
+        data=Artifact(
+            type=ArtifactType.DOCUMENT,
+            created_at=int(time.time()),
+            data=DocumentArtifactData(
+                title="My cooking recipes",
+                content=content,
+                type="markdown",
+            ),
+        ),
+    )
+)
+```
+
+This will send the artifact to the LlamaIndex Server UI, where it will be rendered in the Canvas panel by a renderer depending on the artifact type. For `ArtifactType.DOCUMENT`, this uses a `DocumentArtifactViewer`.
+
+## Available Artifact Types
+
+LlamaIndex Server currently supports the following artifact types:
+
+- `ArtifactType.DOCUMENT`: For text-based documents like Markdown or HTML.
+  - `data` should be an instance of `DocumentArtifactData` which includes `title: str`, `content: str`, and `type: Literal["markdown", "html"]`.
+- `ArtifactType.CODE`: For code snippets.
+  - `data` should be an instance of `CodeArtifactData` which includes `title: str`, `code: str`, and `language: str`.
+
+Ensure you provide the correct data model corresponding to the `ArtifactType` you are sending. You can find these data models in `llama_index.server.api.models`. 
@@ -0,0 +1,38 @@
+# A simple chat app
+
+This guide explains how to set up and use the LlamaIndex server with a simple chatbot agent.
+
+## Prerequisites
+
+- [uv](https://github.com/astral-sh/uv) installed (a fast Python package manager and runner)
+- An OpenAI API key
+
+## Steps
+
+1. **Set the OpenAI API Key**
+
+   Export your OpenAI API key as an environment variable:
+
+   ```sh
+   export OPENAI_API_KEY=your_openai_api_key_here
+   ```
+
+2. **Run the Server Using uv**
+
+   Start the server with the following command:
+
+   ```sh
+   uv run workflow.py
+   ```
+
+   This will launch the FastAPI server using the workflow defined in `main.py`.
+
+3. **Access the Application**
+
+   Open your browser and go to:
+
+   ```
+   http://localhost:8000
+   ```
+
+   You will see the LlamaIndex chat app UI, where you can interact with the agent.
@@ -0,0 +1,14 @@
+from typing import Optional
+
+from llama_index.core.agent.workflow import AgentWorkflow
+from llama_index.core.settings import Settings
+from llama_index.llms.openai import OpenAI
+from llama_index.server.api.models import ChatRequest
+
+
+def create_workflow(chat_request: Optional[ChatRequest] = None) -> AgentWorkflow:
+    return AgentWorkflow.from_tools_or_functions(
+        tools_or_functions=[],
+        llm=Settings.llm or OpenAI(model="gpt-4o-mini"),
+        system_prompt="You are a helpful assistant that can tell a joke about Llama.",
+    )
@@ -0,0 +1,29 @@
+from app.workflow import create_workflow
+from fastapi import FastAPI
+
+from llama_index.server import LlamaIndexServer, UIConfig
+
+
+def create_app() -> FastAPI:
+    app = LlamaIndexServer(
+        workflow_factory=create_workflow,
+        ui_config=UIConfig(
+            app_title="Artifact",
+            starter_questions=[
+                "Tell me a funny joke.",
+                "Tell me some jokes about AI.",
+            ],
+            component_dir="components",
+            dev_mode=True,  # To show the dev UI, should disable this in production
+        ),
+    )
+    return app
+
+
+app = create_app()
+
+
+if __name__ == "__main__":
+    import uvicorn
+
+    uvicorn.run("main:app", host="0.0.0.0", port=8000, reload=True)
@@ -1,4 +1,9 @@
 from llama_index.server.api.routers.chat import chat_router
 from llama_index.server.api.routers.ui import custom_components_router
+from llama_index.server.api.routers.dev import dev_router

-__all__ = ["chat_router", "custom_components_router"]
+__all__ = [
+    "chat_router",
+    "custom_components_router",
+    "dev_router",
+]
@@ -0,0 +1,102 @@
+import os
+import tempfile
+
+from fastapi import APIRouter, HTTPException
+from pydantic import BaseModel, Field
+
+from llama_index.server.settings import server_settings
+from llama_index.server.utils.workflow_validation import validate_workflow_file
+
+
+class WorkflowFile(BaseModel):
+    last_modified: int
+    file_path: str = Field(
+        default="app/workflow.py",
+        description="Relative path to the workflow file",
+    )
+    content: str
+
+
+class WorkflowFileUpdate(BaseModel):
+    content: str
+    file_path: str = Field(
+        default="app/workflow.py",
+        description="Relative path to the workflow file",
+    )
+
+
+class WorkflowValidationResult(BaseModel):
+    valid: bool
+    error: str
+
+
+def dev_router() -> APIRouter:
+    # Use a prefix here to avoid conflicts with other routers
+    # but we probably don't need to do this
+    router = APIRouter(prefix="/dev", tags=["dev"])
+
+    default_workflow_file_path = "app/workflow.py"
+
+    @router.get("/files/workflow")
+    async def get_workflow_file() -> WorkflowFile:
+        """
+        Fetch the current workflow code
+        """
+        # Check if the file exists
+        if not os.path.exists(default_workflow_file_path):
+            raise HTTPException(
+                status_code=400,
+                detail="Dev mode is currently in beta. It only supports updating workflow file at 'app/workflow.py'",
+            )
+        stat = os.stat(default_workflow_file_path)
+        with open(default_workflow_file_path, "r") as f:
+            return WorkflowFile(
+                last_modified=int(stat.st_mtime),
+                file_path=default_workflow_file_path,
+                content=f.read(),
+            )
+
+    @router.post("/files/workflow/validate")
+    async def validate_workflow(file: WorkflowFileUpdate) -> WorkflowValidationResult:
+        """
+        Validate the current workflow code
+        """
+        try:
+            if file.file_path != default_workflow_file_path:
+                raise HTTPException(
+                    status_code=400, detail=f"Updating {file.file_path} is not allowed"
+                )
+            validate_workflow_file(
+                workflow_content=file.content,
+                factory_signature=server_settings.workflow_factory_signature,
+            )
+            return WorkflowValidationResult(valid=True, error="")
+        except Exception as e:
+            raise HTTPException(status_code=400, detail=str(e))
+
+    @router.put("/files/workflow")
+    async def put_workflow_file(update: WorkflowFileUpdate) -> None:
+        """
+        Update the current workflow code
+        """
+        # Validations
+        if update.file_path != default_workflow_file_path:
+            raise HTTPException(
+                status_code=400, detail=f"Updating {update.file_path} is not allowed"
+            )
+        with tempfile.NamedTemporaryFile("w", suffix=".py", delete=False) as tmp:
+            tmp.write(update.content)
+            tmp_path = tmp.name
+        try:
+            # Validate workflow file using the actual callable name from the workflow_factory
+            factory_func_name = server_settings.workflow_factory_signature
+            validate_workflow_file(
+                workflow_path=tmp_path, factory_signature=factory_func_name
+            )
+        except Exception as e:
+            raise HTTPException(status_code=400, detail=str(e))
+        # If all checks pass, overwrite the real file
+        with open(default_workflow_file_path, "w") as f:
+            f.write(update.content)
+
+    return router
@@ -5,7 +5,7 @@ from typing import Optional

 import requests

-CHAT_UI_VERSION = "0.1.6"
+CHAT_UI_VERSION = "0.2.1"


 def download_chat_ui(
@@ -10,7 +10,11 @@ from fastapi.staticfiles import StaticFiles
 from pydantic import BaseModel, Field

 from llama_index.core.workflow import Workflow
-from llama_index.server.api.routers import chat_router, custom_components_router
+from llama_index.server.api.routers import (
+    chat_router,
+    custom_components_router,
+    dev_router,
+)
 from llama_index.server.chat_ui import download_chat_ui
 from llama_index.server.settings import server_settings

@@ -33,6 +37,9 @@ class UIConfig(BaseModel):
    component_dir: Optional[str] = Field(
        default=None, description="The directory to custom UI components code"
    )
+    dev_mode: bool = Field(
+        default=False, description="Whether to enable the UI dev mode"
+    )

    def get_config_content(self) -> str:
        return json.dumps(
@@ -46,6 +53,7 @@ class UIConfig(BaseModel):
                "COMPONENTS_API": f"{server_settings.api_url}/components"
                if self.component_dir
                else None,
+                "DEV_MODE": self.dev_mode,
            },
            indent=2,
        )
@@ -100,24 +108,33 @@ class LlamaIndexServer(FastAPI):
            server_settings.set_url(server_url)
        if api_prefix:
            server_settings.set_api_prefix(api_prefix)
-
-        if self.use_default_routers:
-            self.add_default_routers()
+        server_settings.set_workflow_factory(workflow_factory.__name__)

        if str(env).lower() == "dev":
            self.allow_cors("*")
            if self.ui_config.enabled is None:
                self.ui_config.enabled = True
-
+        else:
+            if self.ui_config.enabled and self.ui_config.dev_mode:
+                raise ValueError(
+                    "UI dev mode requires the environment variable for LlamaIndexServer to be set to 'dev' and start the FastAPI app in dev mode."
+                )
        if self.ui_config.enabled is None:
            self.ui_config.enabled = False

+        # Routers
+        if self.use_default_routers:
+            self.add_default_routers()
+
+        # Should mount ui at the end
        if self.ui_config.enabled:
            self.mount_ui()

    # Default routers
    def add_default_routers(self) -> None:
        self.add_chat_router()
+        if self.ui_config.enabled and self.ui_config.dev_mode:
+            self.include_router(dev_router(), prefix=server_settings.api_prefix)
        self.mount_data_dir()
        self.mount_output_dir()

@@ -11,6 +11,10 @@ class ServerSettings(BaseSettings):
        default="/api",
        description="The prefix for the API endpoints",
    )
+    workflow_factory_signature: str = Field(
+        default="",
+        description="The signature of the workflow factory function",
+    )

    @property
    def file_server_url_prefix(self) -> str:
@@ -40,6 +44,9 @@ class ServerSettings(BaseSettings):
        self.api_prefix = v
        self.validate_api_prefix(v)  # type: ignore

+    def set_workflow_factory(self, v: str) -> None:
+        self.workflow_factory_signature = v
+
    class Config:
        env_file_encoding = "utf-8"

@@ -0,0 +1,54 @@
+"""
+Utilities for validating workflow.py files (syntax and import checks only).
+"""
+
+import ast
+import importlib.util
+from typing import Optional
+
+
+def validate_workflow_file(
+    workflow_path: Optional[str] = None,
+    workflow_content: Optional[str] = None,
+    factory_signature: Optional[str] = None,
+) -> None:
+    """
+    Validate that the workflow file is syntactically correct, can be imported, and defines a callable factory function with the given name.
+    Raises an exception if invalid.
+    """
+    if workflow_path is None and workflow_content is None:
+        raise ValueError("Either workflow_path or workflow_content must be provided")
+
+    # 1. Syntax check
+    if workflow_path is not None:
+        with open(workflow_path, "r") as f:
+            content = f.read()
+    else:
+        if workflow_content is None:
+            raise ValueError(
+                "workflow_content must be provided if workflow_path is not specified"
+            )
+        content = workflow_content
+    try:
+        ast.parse(content)
+    except SyntaxError as e:
+        raise ValueError(f"Syntax error in workflow: {e}")
+
+    # 2. Import check (will catch missing modules, etc.)
+    spec = importlib.util.spec_from_file_location("workflow", workflow_path)
+    if spec is None or spec.loader is None:
+        raise ValueError(f"Could not load module specification for {workflow_path}")
+
+    mod = importlib.util.module_from_spec(spec)
+    try:
+        spec.loader.exec_module(mod)
+    except Exception as e:
+        raise ValueError(f"Import error: {e}")
+
+    # 3. Contract validation: require the given factory function name
+    if factory_signature:
+        if not hasattr(mod, factory_signature):
+            raise ValueError(f"Missing required function: '{factory_signature}'")
+        obj = getattr(mod, factory_signature)
+        if not callable(obj):
+            raise ValueError(f"'{factory_signature}' is not callable")
@@ -1,6 +1,6 @@
 [project]
 name = "llama-index-server"
-version = "0.1.15"
+version = "0.1.16"
 description = "llama-index fastapi server"
 readme = "README.md"
 license = "MIT"
Author	SHA1	Message	Date
github-actions[bot]	bbae802bed	Release 0.2.2 (#638 ) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>	2025-05-22 17:17:34 +07:00
Thuc Pham	25fba4381b	refactor: migrate to Nextjs Route Handler (#625 )	2025-05-22 11:47:24 +07:00
Huu Le	d0618fa2fa	add changeset (#639 )	2025-05-21 14:31:41 +07:00
Huu Le	f3fe3ffc9b	fix: llamacloud generate not working and re-add tests (#636 )	2025-05-21 12:49:44 +07:00
Thuc Pham	6f75d4ab6e	fix: unsupported language in code gen workflow (#633 )	2025-05-21 12:31:11 +07:00
Huu Le	3242738fe4	chore: Fix Python e2e tests (#632 )	2025-05-21 11:30:02 +07:00
Sourabh Kondapaka	17538eb0dd	Fixed bug when traceloop observability is chosen but does not install the latest version (#603 ) Co-authored-by: Marcus Schiesser <mail@marcusschiesser.de>	2025-05-20 11:48:32 +07:00
github-actions[bot]	d3772cb4a2	Release 0.5.15 (#629 ) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>	2025-05-16 16:33:35 +07:00
Huu Le	527075c086	enable dev mode that allows updating code directly in the UI (#624 ) * Enable dev mode that allows updating code directly in the UI * bump server packages	2025-05-16 16:05:56 +07:00
github-actions[bot]	fb7d4da149	chore(release): bump llama-index-server version to 0.1.16 (#587 ) Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>	2025-05-16 15:16:57 +07:00
leehuwuj	5c35b194bb	bump chat ui version	2025-05-16 14:53:57 +07:00
github-actions[bot]	85e5e7e662	Release 0.5.14 (#608 ) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>	2025-05-16 14:41:46 +07:00
Huu Le	58362542c0	chore: add workflow contract for server (#623 )	2025-05-16 14:26:24 +07:00
Thuc Pham	6f44185f68	fix: init messages memory in start event handler (#627 )	2025-05-16 12:45:35 +07:00
Thuc Pham	afe9e9fc16	fix: nodemon should ignore temp file (#622 )	2025-05-15 15:33:24 +07:00
Thuc Pham	1b5a519f13	chore: improve dev experience with nodemon (#621 )	2025-05-15 15:18:12 +07:00
Huu Le	f072308d03	feat: Add dev mode (#610 ) * Add UI components and static assets for chat interface * feat: Add simple chat app example with FastAPI integration * fix: update default workflow file path and improve error handling * update doc * change to file_path * include changes from #614 * fix mypy * support devmode for backend ts server * Revert "support devmode for backend ts server" This reverts commit `bd943fd8c1`. * fix: polling should work when server not yet started * bump chat-ui to fix syntax highlight issue * fix: missing language for code editor * enhance UI with shadow overlay * enhance doc * fix minor UI bugs * enhance doc * remove unessesary debug log * fix wrong check * increase delay time before trigger polling * feat: support dev mode for backend ts server (#616) * feat: support dev mode for backend ts server * update message * validate typescript file * fix: format * use temp file to avoid server restart * fix format * use npx tsc * remove typescript deps --------- Co-authored-by: thucpn <thucsh2@gmail.com> Co-authored-by: Thuc Pham <51660321+thucpn@users.noreply.github.com>	2025-05-15 14:56:08 +07:00
Huu Le	1df8cfbdc2	refactor: split artifacts use case into document generator and code generator (#617 ) * split artifacts use case to code generator and document generator * add changeset * fix package version * fix typing * bump openai * fix package * fix typing * fix: improve type handling and clean up UI event component - Removed unnecessary string conversion for userInput in code_generator and deep_research workflows. - Updated userRequest type to MessageContent for better type safety. - Cleaned up the UI event component by removing redundant indicatorClassName logic. * docs: word smith * better handler typing * refactor: remove redundant UI event handling in workflows --------- Co-authored-by: Marcus Schiesser <mail@marcusschiesser.de>	2025-05-15 14:22:21 +07:00