PAL MCP Server

PAL MCP Server connects coding assistants to other AI models for second opinions, code review and planning. It's for developers who want to consult several models within the same coding session, using clients such as Claude Code, Codex CLI, Gemini CLI or Cursor.

• Model collaboration: Compare recommendations, debate approaches and carry conversation context between models and tools without starting each consultation from scratch.
• Code review and debugging: Request reviews with severity-ranked findings, investigate possible causes of bugs and check proposed changes before committing.
• CLI subagents: The clink tool delegates work to separate Claude Code, Codex CLI or Gemini CLI sessions with roles such as planner or code reviewer, returning their final results to the main session.
• Planning and development: Break larger projects into actionable tasks and use additional tools for refactoring, test generation, security audits and documentation.
• Images and API research: Consult vision-capable models about screenshots or diagrams, and request current API and SDK documentation lookups.

The server runs locally and supports Ollama for local Llama or Mistral models. Those model requests can stay on your machine; cloud requests send their supplied context to providers such as Gemini, OpenAI, Anthropic, Azure, X.AI, OpenRouter or DIAL and require internet access and provider credentials. Local model support doesn't make a connected cloud coding assistant offline. Some development tools, including refactoring, test generation and security audits, are disabled by default.