Chain of Draft

Enables iterative reasoning through structured drafts with explicit reasoning chains, allowing for focused critiques and targeted revisions to improve problem-solving quality through systematic refinement.

MCP Chain of Draft Server provides a tool for enhancing reasoning through iterative critique and revision cycles. Developed by bsmi021, this TypeScript implementation exposes a Chain of Draft tool that structures thinking into drafts with explicit reasoning chains, allowing for focused critiques and targeted revisions. The server maintains session state to track thought history and branching paths, making it particularly valuable for complex problem-solving tasks where initial reasoning may contain errors or gaps. It integrates with the Model Context Protocol SDK to enable AI assistants to improve their reasoning quality through systematic refinement rather than single-pass thinking.