Sequential Thinking MCP Server

This repository contains a Model Context Protocol (MCP) server that provides a sequentialthinking tool. The tool facilitates dynamic and reflective problem-solving through a chain of thoughts, allowing AI agents to break down complex problems, revise past thoughts, and explore branches of logic before arriving at a conclusion.

This project is configured to run in two modes:

1. Stdio: Standard MCP transport for local IDE and CLI integrations.
2. Server-Sent Events (SSE): Deployed as a secure, remote HTTP gateway on Google Cloud Run.

Local Development

Install dependencies and build the TypeScript code:

npm install
npm run build

To run the Stdio server locally:

node dist/index.js

Cloud Run Deployment (SSE Gateway)

This project can be easily cloned and deployed to any Google Cloud account or project. This allows anyone to host their own secure, remote Sequential Thinking MCP server.

Prerequisites

1. Install the Google Cloud CLI (gcloud).
2. Authenticate with your Google account:

   gcloud auth login
   

3. Set your active Google Cloud Project (ensure billing is enabled):

   gcloud config set project YOUR_PROJECT_ID
   

4. Enable the required APIs for your project:

   gcloud services enable run.googleapis.com cloudbuild.googleapis.com
   

Deploying

Once authenticated, you can deploy the service using the provided Makefile:

make deploy

What this does:

• Packages the application and builds a Docker container using the provided Dockerfile.
• Generates a secure, random MCP_API_KEY (if one is not already provided).
• Deploys the container to Google Cloud Run (default region: us-central1).
• Outputs the Service URL and your secure API key.

Client Configuration

Once deployed, you can configure your MCP client (Cursor, Hermes, Antigravity, Claude Desktop, etc.) to connect to the SSE gateway.

Provide your client with:

• URL: <YOUR_CLOUD_RUN_URL>/sse
• Headers: Authorization: Bearer <YOUR_MCP_API_KEY>