agentconnect is an open-source platform that brings multiple AI agents into a team's chat, repository, and task workflows. People and agents collaborate in shared conversations, with work triggered by messages, issues, pull requests, webhooks, or schedules.
← Back to Category
AI Coding
qwen-code
qwen-code is an open-source AI coding agent available through a terminal, editors, desktop app, browser and chat channels. It supports multiple model APIs, with headless execution and SDKs for development automation.
Directory information is for reference. Check the official site for current features, pricing and availability. Listing does not imply hands-on testing or a purchase recommendation.
About this product
Both the framework and Qwen models are open source. A persistent daemon lets clients share agent sessions across multiple workspaces and reconnect to active sessions.
Key Features
- Supports subagents, agent teams, automatic memory, automatic skills and MCP.
- Connects to OpenAI, Anthropic, Gemini and Qwen APIs, plus local models through Ollama or vLLM, with runtime switching.
- Offers VS Code, Zed and JetBrains integrations, alongside desktop apps for macOS, Windows and Linux.
- Connects to Telegram, DingTalk, WeChat and Feishu.
- Runs headlessly in scripts, CI/CD pipelines and batch jobs.
- Provides TypeScript, Python and Java SDKs.
- Includes built-in skills for reviews, batch tasks, loops and bug fixes.
Use cases
- Explain a repository's structure and identify where to begin reading.
- Run agent tasks from CI/CD pipelines or batch scripts.
- Share agent sessions among browser, editor and SDK clients.
Who it is for
- Developers who want an AI agent in their terminal or editor.
- Teams integrating coding agents into automated workflows or custom applications.
Requirements
- NPM installation requires Node.js 22 or later.
- Authentication requires a Qwen Plus or Pro subscription, or a configured provider API key.
Quick Start
- Install through NPM with npm install -g @qwen-code/qwen-code@latest.
- Restart the terminal after installation so environment variables take effect.
- Open the project directory in a terminal and run qwen.
- Run /auth in the session and sign in with a Qwen subscription or configure an API key.
- Ask the agent to explain the repository and suggest where to start.
Limitations and considerations
- The Web UI and daemon interface are marked experimental.
- The daemon defaults to loopback access without bearer authentication; local processes that can reach it can use the full API, including code execution.
License and links
Code license: Apache-2.0
Pricing notes
- Qwen account authentication requires a Plus or Pro subscription; provider API-key configuration is also supported.
Privacy and data
- The README links to a Terms of Service and Privacy Notice covering Chrome extension and Browser Use data handling.
Sources
Source review: 2026-11-08
Related Tools
AI Coding
SUZENT is an open-source, local-first personal AI agent that keeps memory in files you control. You can set rules for its actions, choose cloud or local models, and have it follow up on projects while you’re away.
GitHub Copilot is an AI assistant for software development. It can suggest code, explain repositories, edit files, review changes, and work on assigned tasks that developers can inspect as pull requests.
Cursor is a desktop development tool with an AI coding agent. It can read project files, explain a codebase, make requested edits, and support a workflow of reviewing diffs and running project checks.
v0 is Vercel’s AI application builder. It turns natural-language ideas, wireframes, or mockups into web apps that can be refined through previews, code views, and design controls, connected to services, and published.
The former Codeium website now points to Devin Desktop, the renamed Windsurf IDE. It combines a code editor with local and cloud agent management for organizing tasks, reviewing progress and inspecting changes.