Qwen Gate

What does this agent do, and when should you use it?

Qwen Gate is a self-hosted gateway that signs in to chat.qwen.ai with browser automation and exposes Qwen models through OpenAI-compatible endpoints. It runs TypeScript directly with Bun and provides `/v1/chat/completions` and `/v1/models` for clients configured with the local service URL. Features include SSE streaming, account rotation, a reusable session pool, and parsing of tool calls. A web dashboard covers request activity, accounts, network debugging, and settings. Playwright is used for login and authentication, while API requests use the wreq-js Node.js fetch transport. It fits developers who can operate the service and manage Qwen accounts, and whose clients speak the OpenAI API format; the project says it is for educational and study purposes and is not affiliated with Qwen or Alibaba.

Run qg or bun start to launch the Hono API and dashboard on the configured port (26405 by default). An administrator adds Qwen email and password credentials at /dashboard/accounts; the gateway uses Playwright to sign in and persist sessions. A client then sends model, message, and optional tool definitions to /v1/chat/completions, or queries /v1/models. The gateway obtains responses from Qwen, can stream them over SSE, and manages account rotation, cooldowns after rate limits, and retries; text-form tool calls are parsed into OpenAI-style tool call objects. The dashboard includes overview, logs, accounts, network debugging, and settings pages, with configuration and request logging behavior controlled by settings.

1. A Claude Code, Cursor, or OpenCode user wants to point an existing OpenAI-compatible client at a self-hosted endpoint for Qwen models.
2. A developer managing several Qwen accounts wants round-robin distribution, automatic failover, and cooldown tracking.
3. A developer whose client expects SSE wants to route streaming requests through Qwen web sessions.
4. An operator wants a local dashboard for request logs, account status, and outbound network inspection.
5. A developer wants to experiment with tool calls and accepts that the gateway parses them from generated text with documented limitations.

How do you install or deploy this agent?

On Linux or macOS, the install script clones the repository, installs dependencies, and creates CLI aliases; manual setup requires Git and Bun. On Windows, use the PowerShell installer or clone the repository and run bun install. You will need a Qwen account and must add it in the dashboard. The README specifies Bun 1.3+ and uses browser automation for authentication.

curl -sSL https://raw.githubusercontent.com/youssefvdel/qwen-gate/main/install.sh | bash
cd qwen-gate
qg

For Windows, the documented installer command is:

powershell -ExecutionPolicy Bypass -c "curl.exe -sSL https://raw.githubusercontent.com/youssefvdel/qwen-gate/main/install.ps1 | iex"

How do you use this agent?

Start the service, then open the dashboard and add a Qwen account. Set API_KEY in config.json if you want API authentication; its default is empty, which disables auth. Configure a client with base URL http://localhost:26405/v1 and send a Bearer token if one is configured. The example below starts the server and makes a first chat completion request:

qg

curl -X POST http://localhost:26405/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-api-key" \
  -d '{
    "model": "qwen3-max",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Replace your-api-key with the value configured in config.json; omit authorization when API auth is disabled. Set "stream": true to request SSE output.

What are this agent's strengths and limitations?

How does this agent compare with similar options?

Key facts side by side with the most closely related agents.

How does FollowAgents rate this agent?

FAQ