Retrieval-Augmented Thinking MCP Server

A server implementing Chain of Draft reasoning for enhanced problem-solving capabilities using OpenAI.

Retrieval Augmented Thinking MCP Server - A reasoning tool that processes structured thoughts with metrics, branching, and revision capabilities.

git clone https://github.com/stat-guy/retrieval-augmented-thinking.git cd retrieval-augmented-thinking npm install -g .

Success indicator:If you seeRAT MCP Server (Node.js) running on stdio, your installation is ready!

Add to your Claude Desktop configuration file:

macOS:~/Library/Application Support/Claude/claude_desktop_config.json
Windows:%APPDATA%/Claude/claude_desktop_config.json

{ "mcpServers": { "retrieval-augmented-thinking": { "command": "npx", "args": ["mcp-server-rat-node"] } } }

After adding the configuration, restart Claude Desktop to load the RAT server.

The server provides a singlerattool for processing structured thoughts:

// Basic usage { "thought": "I need to analyze this problem step by step...", "nextThoughtNeeded": true, "thoughtNumber": 1, "totalThoughts": 3 } // With revision { "thought": "Let me reconsider my previous analysis...", "nextThoughtNeeded": false, "thoughtNumber": 2, "totalThoughts": 3, "isRevision": true, "revisesThought": 1 } // With branching { "thought": "Alternative approach: what if we consider...", "nextThoughtNeeded": true, "thoughtNumber": 2, "totalThoughts": 4, "branchFromThought": 1, "branchId": "alt-path-1" }

- thought(string): The thought content to process
- nextThoughtNeeded(boolean): Whether another thought is needed to continue
- thoughtNumber(integer): Current thought number in the sequence
- totalThoughts(integer): Total expected thoughts (adjustable)

- isRevision(boolean): Whether this revises a previous thought
- revisesThought(integer): The thought number being revised
- branchFromThought(integer): Thought number to branch from
- branchId(string): Unique identifier for this branch
- needsMoreThoughts(boolean): Extend beyond totalThoughts if needed

{ "thought_number": 1, "total_thoughts": 3, "metrics": { "complexity": 0.342, "depth": 0.521, "quality": 0.643, "impact": 0.289, "confidence": 0.758 }, "analytics": { "total_thoughts": 5, "average_quality": 0.612, "chain_effectiveness": 0.145 }, "next_thought_needed": true, "visual_output": "┌─ 💭 Thought 1/3 ─────────────────┐\\n│ Analysis shows clear patterns... │\\n├─ Metrics ──────────────────────┤\\n│ Quality: 0.64 | Impact: 0.29... │\\n└─────────────────────────────────┘" }

The installation process includes automatic permission fixes. If you encounter issues:
- Ensure you have Node.js and npm installed
- Try the alternative installation method:

npm install -g git+https://github.com/stat-guy/retrieval-augmented-thinking.git

chmod +x $(npm bin -g)/mcp-server-rat-node

Before configuring Claude Desktop, always verify:

If this shows "RAT MCP Server (Node.js) running on stdio", you're ready to configure Claude Desktop.

This MCP server is licensed under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For more details, please see the LICENSE file in the project repository.

An MCP server to interact with OpenAI's ChatGPT API for conversational AI and text generation.

An AI-powered email intelligence platform that integrates with Gmail and OpenAI. It can be run as a CLI tool or deployed on AWS Lambda for enhanced capabilities.

Persistent memory for any AI assistant. Zero token cost until recall. Stores memories in local SQLite, ranks by 6-factor scoring, returns results 79% smaller than JSON. Works with Claude, ChatGPT, Grok, Cursor, Windsurf, and any MCP client.

An MCP server which brings Jotform to your AI client or LLM

After Effects MCP is a full-featured automation bridge that connects AI clients (like VS Code, Claude Desktop, and Claude Code) to Adobe After Effects through MCP, enabling scripted control of compositions, layers, effects, keyframes/graph easing, presets, markers, audio levels, waveform analysis, and effect discovery via a live bridge panel.

Project management your AI can actually run — connect Claude, ChatGPT, Cursor & Codex to one board over MCP.

Compare LLM cost & latency on one prompt, translate PDF keeping layout, cited research, make PPTX

AIOProductOS spine over MCP — customers, revenue, feedback, work, analytics on one typed record.

The memory layer for AI coding tools. Local-first, semantic, 9 MCP tools with consolidation and project scoping. Works with Claude Code, Cursor, Windsurf & any MCP client.

One MCP server for Claude, ChatGPT & Gemini — wraps your ERPs, CRMs, APIs and knowledge base into a single governed endpoint.