thought-chain
Thought Chain MCP - A Model Context Protocol server for implementing thought chain reasoning and sequential processing
claude mcp add --transport stdio cbuntingde-thought-chain-mcp npx thought-chain-mcp
How to use
The Thought Chain MCP Server turns any AI model into an advanced reasoning engine by providing a structured, persistent framework for step-by-step thinking. It lets you build, save, and recall thought chains that decompose problems into logical steps, with reflections and cross-session continuity backed by an SQLite database. This enables models that may struggle with multi-step reasoning to produce organized workflows, maintain context over conversations, and generate auditable decision trails across different agents or editors. The server is designed to be universally compatible with MCP-enabled tools, including Claude Desktop, Cursor, Cline, Roo Code, Windsurf, and VS Code MCP extensions, ensuring you can reuse your thought chains across environments.
To use it, install the server via NPX (recommended) or run it directly through your preferred MCP-enabled editor or tool. Once running, configure your editor or assistant to point at the thought-chain MCP configuration, typically by invoking npx thought-chain-mcp (or its global equivalent). As you interact with the model, the Thought Chain server will store chains, steps, reflections, and outcomes in a local SQLite database, enabling you to recall or search previous reasoning across sessions and devices.
How to install
Prerequisites:
- Node.js 20.0.0 or higher
- npm (latest stable)
- An MCP-compatible AI coding assistant or editor
Installation options:
Option A: NPX Installation (Recommended)
- Ensure Node.js and npm are installed
- Run:
npx thought-chain-mcp
Optionally enable global access:
npm install -g thought-chain-mcp
Then you can start the server with:
thought-chain-mcp
Option B: Direct NPX Usage (No Installation)
- Run directly without installing:
npx thought-chain-mcp@latest
Option C: Local Development Setup
- Navigate to the project directory:
cd /path/to/mcpservers/thought-chain-mcp
- Install dependencies:
npm install
- Configure your MCP tool to use the server, as shown in the Quick Start examples within the README (examples for Claude Desktop, Cursor, Cline, Roo Code, Windsurf, and generic MCP configurations).
Database note: On first run, the server will create a SQLite database at your home directory at ~/.thought-chain-mcp/thoughts.db. This stores all thought chains locally and persists across installations.
Additional notes
Tips and common issues:
- Default database location: ~/.thought-chain-mcp/thoughts.db. You can safely delete the directory to start fresh.
- If you upgrade or reinstall, ensure your MCP configuration points to the same server command (npx thought-chain-mcp) or its global equivalent.
- When using NPX, ensure you have network access to fetch the package; locally installed setups may offer faster startup.
- The server supports multiple editors/tools; you only need to configure the relevant tool with the thought-chain configuration shown in the README (each tool has examples).
- For cross-session continuity, avoid altering the database path mid-use unless migrating data intentionally.
- If you encounter permissions issues on Linux/macOS when creating ~/.thought-chain-mcp, ensure your user has write access to your home directory.
Related MCP Servers
after-effects
MCP Server for Adobe After Effects. Enables remote control (compositions, text, shapes, solids, properties) via the Model Context Protocol using ExtendScript.
mongodb-lens
🍃🔎 MongoDB Lens: Full Featured MCP Server for MongoDB Databases
crash
MCP server for structured and efficient reasoning with step validation, branching, and revisions.
grumpydev
Let the grumpy senior dev review your code with this MCP server
short-url
简单易用的短链接生成工具,完全开源、免费、无需登录,可私有化部署,链接永久有效!
civitai
A Model Context Protocol server for browsing and discovering AI models on Civitai