# Qwen MCP Server

> Qwen AI Chat Guide for Qwen3.6-Plus, Flash, and 1M Context

[![MCP Badge](https://lobehub.com/badge/mcp/rocnubie-chat-mcp)](https://lobehub.com/mcp/rocnubie-chat-mcp)
[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](./LICENSE)
[![Zero Config](https://img.shields.io/badge/setup-zero--config-7c3aed)](#installation)
[![Node](https://img.shields.io/badge/node-%3E%3D18-339933?logo=node.js&logoColor=white)](https://nodejs.org)
[![MCP](https://img.shields.io/badge/MCP-1.0-blue)](https://modelcontextprotocol.io)
[![smithery](https://smithery.ai/badge/chat)](https://smithery.ai)

<p align="center"><a href="https://qwenchat.online"><img src="./assets/hero.svg" alt="Qwen" width="720" /></a></p>

A Model Context Protocol server that exposes the canonical Qwen knowledge surface — models, prompts, and chat workflows, pricing, FAQ, official links — to MCP-compatible AI clients such as Claude Desktop, Cursor, Windsurf, and Continue. Read-only, no API keys, no quota, ~50 ms cold start.

Official website: https://qwenchat.online

## 💬 About Qwen

Qwen Chat (qwenchat.online) is a browser-based workspace for interacting with Alibaba's Qwen3 family of language models. The site provides direct chat access alongside a structured evaluation environment where users can test model capabilities across distinct task categories — code, documents, visual inputs, and web generation — without setting up local infrastructure. It is designed for people who want to move quickly from curiosity to practical assessment: load a file, paste a prompt, and observe how the model handles real work rather than toy examples.

## Key Features

- **1M context window** — carry multiple long files, research documents, or codebases in a single request without manually summarising between steps.
- **Model lane comparison** — switch between Qwen3 Plus and Flash variants through the same interface to compare speed and quality tradeoffs on the same input.
- **OpenAI and Anthropic API compatibility** — the underlying models expose both API surfaces, so existing tooling, SDKs, and prompt templates migrate with minimal changes.
- **Benchmark transparency** — the site surfaces specific third-party scores (Terminal-Bench 2.0, OmniDocBench 1.5) rather than abstract capability claims, letting users calibrate expectations before committing to integration.
- **Multimodal input support** — the chat accepts screenshots, PDFs, and diagrams, enabling visual reasoning tasks such as UI analysis, screen grounding, and document layout parsing.
- **Preset task examples** — the interface ships with ready-made prompts covering code review, document analysis, and web prototyping, reducing the setup time needed to start a meaningful test.

## Use Cases

- **Code and repository work** — reviewing diffs, diagnosing build failures, or walking through a repository repair task using the extended context window to hold multiple files at once.
- **Long-document analysis** — uploading policy documents, research papers, or technical specifications and asking structured questions without hitting context limits.
- **Visual and UI tasks** — pasting screenshots for layout analysis, accessibility review, or screen grounding where the model identifies interactive elements by position.
- **Rapid web prototyping** — generating functional HTML or component scaffolding from a brief description and iterating in the same session.
- **API migration evaluation** — testing whether Qwen models can serve as a drop-in replacement for OpenAI or Anthropic endpoints in an existing workflow before switching production traffic.

## Who Is It For

Qwen Chat is aimed primarily at developers, technical leads, and product teams who are already using language models in their day-to-day work and want to evaluate whether Qwen3 fits their stack. It suits teams that handle long documents or large codebases and need a context window beyond what standard tiers offer. It is also relevant to engineers assessing API compatibility before committing to a migration — the site provides enough structured evidence to make that call without requiring a full integration first. Individual developers building side projects or prototypes who want a capable model accessible directly from the browser will find it practical as well.

## Tools

### `list_models`
Return the canonical list of chat models exposed on the site, with capability notes. (Qwen)

_Input:_ no parameters. _Returns:_ text/markdown.

### `get_pricing`
Return the canonical pricing entry point for Qwen.

_Input:_ no parameters. _Returns:_ text/markdown.

### `get_official_links`
Return the canonical list of official links for Qwen (website, support, docs when available).

_Input:_ no parameters. _Returns:_ text/markdown.

## Resources

- `site://chat/models` — Supported chat models and capability notes.
- `site://chat/pricing` — Canonical pricing entry point.
- `site://chat/faq` — Short FAQ generated from public site metadata.
- `site://chat/links` — Canonical URLs to share with users.

## Prompts

### `tell_me_about_chat`
Summarize what the site is, who it's for, and how it works. — Qwen

### `start_chat_session_chat`
Open a chat-evaluation session against the site's models, with sensible defaults. — Qwen

## Installation

### Install via Smithery

```bash
npx -y @smithery/cli install chat-mcp --client claude
```

(Replace `claude` with `cursor`, `windsurf`, or `continue` for those clients.)

### Install from source

```bash
git clone https://github.com/rocnubie/chat-mcp.git
cd chat-mcp
pnpm install
```

Then add to your MCP client config (`claude_desktop_config.json` for Claude Desktop, `mcp.json` for Cursor / Windsurf / Continue):

```json
{
  "mcpServers": {
    "chat-mcp": {
      "command": "node",
      "args": [
        "/absolute/path/to/chat-mcp/src/index.mjs"
      ]
    }
  }
}
```

### Debug with MCP Inspector

```bash
npx @modelcontextprotocol/inspector node src/index.mjs
```

## Official Links

- Website: https://qwenchat.online
- Pricing: https://qwenchat.online/pricing
- GitHub: https://github.com/Rocniubi/MSA
- Support: support@qwenchat.online

## Development

```bash
pnpm install
pnpm start                 # run the server over stdio
```

## License

MIT