MCP Thought Server
by bsmi021
The MCP Thought Server provides advanced thinking tools via the Model Context Protocol (MCP) to enhance the reasoning, planning, and iterative refinement capabilities of AI agents. It guides AI agents through structured cognitive processes to tackle complex tasks more effectively.
Last updated: N/A
What is MCP Thought Server?
The MCP Thought Server is a Node.js application that provides a suite of tools accessible via the Model Context Protocol (MCP). These tools are designed to aid AI agents in complex problem-solving by offering structured thinking methodologies.
How to use MCP Thought Server?
To use the server, first install the dependencies using npm install. Then, build the project with npm run build and start the server using npm start. The server listens for MCP connections and provides tools that can be accessed by sending requests with specific parameters as defined in the *Params.ts files. Configuration can be adjusted using environment variables.
Key features of MCP Thought Server
Sequential Thinking
Chain of Draft
Integrated Thinking
Set Feature (Debug Tool)
Advanced Confidence Scoring (Semantic Relevance & Optional LLM Coherence Check)
SQLite Persistence for Draft History
Use cases of MCP Thought Server
Guiding AI agents through step-by-step problem-solving
Facilitating iterative content generation and refinement
Combining structured reasoning and iterative content refinement for complex tasks
Debugging and diagnostics of AI agent behavior
Improving the quality and relevance assessment of AI outputs
FAQ from MCP Thought Server
What is the purpose of the thought parameter in sequentialThinking?
What is the purpose of the thought parameter in sequentialThinking?
The thought parameter is required even on the first call (thoughtNumber: 1) and must contain the initial problem statement or analysis.
When can I use category.type: 'final' in integratedThinking?
When can I use category.type: 'final' in integratedThinking?
The category.type: 'final' can only be used when thoughtNumber equals totalThoughts.
How do I configure the SQLite database path?
How do I configure the SQLite database path?
You can configure the SQLite database path using the MCP_SQLITE_PATH environment variable. If not set, it defaults to data/mcp-thought-server.sqlite.
How do I enable the LLM coherence check?
How do I enable the LLM coherence check?
To enable the LLM coherence check, you need to set the COHERENCE_API_KEY and COHERENCE_CHECK_MODEL environment variables. Optionally, you can also set COHERENCE_API_BASE, COHERENCE_HTTP_REFERER, and COHERENCE_X_TITLE.
What should I do if I get an 'Original draft X not found' error?
What should I do if I get an 'Original draft X not found' error?
This indicates an issue retrieving draft history. Ensure the SQLite database file path is correct and the server process has write permissions to that location. Check server logs for database errors.