This project provides a Model Context Protocol (MCP) server that proxies the Todoist REST API v2. It allows AI models (like Claude or ChatGPT) to interact with your Todoist tasks through a standardized interface.
- Node.js 16 or higher
- A Todoist account and API token (get it from Todoist Integrations settings)
- Clone this repository
- Install dependencies:
npm install - Build the project:
npm run build
- Copy
.env.exampleto.env:cp .env.example .env - Edit
.envand replaceYOUR_TODOIST_API_TOKENwith your actual Todoist API token.
Run the server with:
npm start
This will start the MCP server using the included Todoist OpenAPI specification.
The server is designed to work with MCP-compatible AI systems. It provides a standardized interface for models to:
- List available tools (Todoist API endpoints)
- Call specific tools with parameters
- Receive structured responses
This project is configured for deployment on Railway. Just connect your repository to Railway, and it will automatically build and deploy.
Make sure to set the OPENAPI_MCP_HEADERS environment variable in your Railway project settings with your Todoist API token:
{"Authorization":"Bearer YOUR_TODOIST_API_TOKEN"}
You can also run the application using Docker:
-
Build the Docker image:
docker build -t openapi-mcp-server . -
Run the container:
docker run -p 3000:3000 -e OPENAPI_MCP_HEADERS='{"Authorization":"Bearer YOUR_TODOIST_API_TOKEN"}' openapi-mcp-server
If the server fails to start or doesn't respond correctly:
- Check that your Todoist API token is valid and correctly formatted in the environment variable
- Ensure the OpenAPI specification file is valid and accessible
- Look for error messages in the console output
MIT