coleam00
Feature request: Add Qwen Code CLI integration (Claude Code–style)
Feature request: Add Qwen Code CLI integration (Claude Code–style)
Summary Archon positions itself as a command center/MCP server for coding assistants and already highlights connectivity to tools like Claude Code. Please add first-class integration for Qwen Code, a command-line AI workflow tool optimized for Qwen3-Coder, similar in spirit to Claude Code. This would give Archon users a powerful, open, and cost-friendly alternative in the same CLI category. GitHub
Why this matters
Qwen Code is an actively maintained CLI adapted from Gemini CLI and tuned for Qwen3-Coder, with strong code understanding/edit/editing workflows. GitHub
Archon’s README explicitly calls out connecting Claude Code and other assistants; Qwen Code is the closest “sibling” in the CLI space and would fit that pattern. GitHub
Archon already supports multiple LLM providers (OpenAI, Gemini, Ollama). Adding Qwen Code improves user choice, especially for users who prefer self-hostable/open options or Qwen’s OAuth/free tiers. GitHub
Proposed integration (minimal viable)
CLI adapter parity with Claude
Mirror the existing Claude CLI integration (e.g., command templates, project-root context, task routing) with a .qwen/ commands folder or equivalent adapter layer invoking the qwen binary.
Allow launching a Qwen Code session anchored to the current Archon project/workspace (working directory, repo root).
Expose common actions (code explain/refactor/test/doc generation) via Archon UI buttons or quick commands that call qwen with the right prompts.
Config & auth
Document installing Qwen Code and verifying the binary:
npm install -g @qwen-code/qwen-code qwen --version
Document two auth paths:
Qwen OAuth (fast start; generous free quota). Users can just run qwen once and sign in. GitHub
OpenAI-compatible API for ModelStudio/OpenRouter/ModelScope/DashScope with OPENAI_API_KEY, OPENAI_BASE_URL, OPENAI_MODEL. GitHub
UX hooks in Archon
“Connect Qwen Code” card alongside Claude Code/Kiro/Cursor/Windsurf in the Integrations/MCP Dashboard with copy-paste config and one-click test. GitHub
Optional: picker to choose “Assistant: Claude Code | Qwen Code” per project.
(Nice-to-have) Future directions
If/when Qwen Code exposes an MCP client/server path, document connecting it to Archon’s MCP server (default port 8051). Today Archon already advertises MCP-based connectivity; Qwen/Qwen-Agent are strengthening MCP support across the ecosystem, so this may become straightforward. GitHub
Acceptance criteria
Docs page: “Using Qwen Code with Archon” (install, auth, basic flows). GitHub
Example project showing: “Explain codebase → refactor file → generate unit tests” flow via Qwen Code from within Archon. GitHub
Archon UI affordance to start/attach a Qwen Code session in the current project.
Parity with existing Claude CLI command templates (where applicable).
References
Archon README: positioning, MCP role, and mention of Claude Code/Kiro/Cursor/Windsurf. GitHub
Archon: multi-LLM support (OpenAI, Ollama, Gemini). GitHub
Thank you!
Thanks for the request. We will take a look at it and take in consideration for implementing.
I will close this one, Archon works with Qwen Coder through MCP and through Agents.md if you want to use it to contribute
Thanks for the suggestion @undef16
