Skip to content

sdk

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History
 
 

sdk

README.md

Qwen Code SDK Integration Demos

This repo demonstrates how to build AI applications using @qwen-code/sdk. It contains two demos:

  • CLI demo: an interactive terminal chatbot
  • Next.js demo: a web chat app with streaming responses (SSE)

Prerequisites

1) Install Qwen Code CLI (global)

The SDK expects a qwen executable to be available on your machine.

After installation, verify it works:

qwen --version

2) Install dependencies

Install project dependencies (this repo uses @qwen-code/sdk@^0.1.1 as defined in package.json):

npm install

CLI demo

Features

The interactive CLI chatbot (demo/cli-chatbot.ts) demonstrates:

  • Multi-turn conversation via an async generator that continuously yields user messages
  • Streaming output by printing partial deltas as they arrive
  • Tool call display by printing tool-use blocks and tool results
  • Tool permission control via the canUseTool callback

Run the CLI demo

npm run start:cli

Usage

  1. Start the program and you will see a prompt
  2. Type a message and press Enter
  3. The assistant responds in a streaming fashion
  4. If tools are used, the tool calls/results are printed
  5. Type quit, exit, or q to exit

Core idea

The CLI demo builds an async iterable of SDKUserMessage and passes it to the SDK query() API, then consumes the streamed SDKMessage output with for await ... of.

Next.js demo

What it demonstrates

The Next.js 16 web chat app demonstrates:

  • Client-side session id to group turns on the UI
  • Streaming responses using Server-Sent Events (SSE)
  • Multi-turn context by sending history and building a single prompt string server-side
  • Error handling for network/stream failures

Run the Next.js demo

Development:

npm run dev

Production:

npm run build npm start

The app runs at http://localhost:3000.

API endpoints

POST /api/chat/init

Creates a new session id (demo mode; no server-side persistence).

Response example:

{ "sessionId": "550e8400-e29b-41d4-a716-446655440000", "status": "success", "message": "Session created successfully" }

Note: the current UI initializes the session id on the client (see src/app/chat/page.tsx), so calling this endpoint is optional.

POST /api/chat/send

Sends a message and receives a streaming SSE response.

Request body:

{ "sessionId": "550e8400-e29b-41d4-a716-446655440000", "prompt": "Hello, please help me write a React component", "history": [ { "role": "user", "content": "..." }, { "role": "assistant", "content": "..." } ] }

Response:

Returns an SSE stream where each event is a JSON-encoded SDK message (i.e., SDKMessage) in data: ...\n\n format. The server ends the stream for the current turn when it sees a type: "result" message.

Project structure

qwen-code-sdk-chat-demo/ ├── demo/ │ └── cli-chatbot.ts # CLI chatbot demo ├── src/ │ ├── app/ │ │ ├── api/ │ │ │ └── chat/ │ │ │ ├── init/ # Session init API │ │ │ └── send/ # SSE streaming chat API │ │ ├── chat/ # Chat page │ │ └── page.tsx # Home page │ ├── components/ │ │ ├── chat/ # Chat UI components │ │ └── ui/ # Base UI components │ ├── lib/ │ │ ├── chat-session.ts # Advanced session manager (not wired by default) │ │ └── utils.ts # Utilities │ └── types/ │ └── chat.ts # App-level types (not required by the demos) ├── package.json └── README.md

Development notes

Requirements

  • Node.js 20+
  • npm (or pnpm)

Scripts

  • npm run dev - start Next.js dev server
  • npm run build - build production bundle
  • npm start - start production server
  • npm run lint - run ESLint
  • npm run start:cli - start the CLI chatbot demo

Tech stack

  • Framework: Next.js 16 (App Router)
  • Language: TypeScript
  • UI: React 19, Tailwind CSS 4
  • SDK: @qwen-code/sdk (this repo uses ^0.1.1)

License

This project is provided by Qwen Team for learning and reference.