basecamp

The Basecamp MCP Server provides a TypeScript-based MCP integration for Basecamp 3, exposing a suite of tools that integrate Basecamp data with Claude Desktop and Cursor IDE via the Model Context Protocol. After you set up and authenticate, you can configure the server to expose 46 tools that cover core Basecamp capabilities, including retrieving projects, lists, and items, managing documents, cards, and workflows, and subscribing to webhooks or events. Once configured, Claude Desktop and Cursor IDE can discover the basecamp MCP server and expose its tools in their respective interfaces, allowing you to interact with Basecamp data directly from chat or editor contexts.

To use the tools, install the server via NPX (npx jhliberty/basecamp-mcp-server), run the setup flow to authenticate with Basecamp, configure the environment for your AI assistant (Claude Desktop or Cursor IDE), and then restart your assistant so it can load the new MCP configuration. In Claude Desktop, you’ll see the basecamp tools listed under the MCP tools area and can enable them; in Cursor, you’ll be able to call individual tools like get_projects, get_todos, create_card, and more. Typical workflows include listing projects, inspecting todo lists, fetching card details, creating or updating cards and documents, and subscribing to or managing webhooks.

If you’re building custom prompts or automations, you can reference the available tool names (e.g., get_projects, get_todos, create_card, update_document) to orchestrate Basecamp data within your AI-assisted workflows.

Prerequisites:

• Node.js 18+ with ES module support
• A Basecamp 3 account and a Basecamp OAuth application (create one at https://launchpad.37signals.com/integrations)
• An environment to run the MCP server (local development or deployment)

Install and configure the server locally:

1. Install dependencies and build (local development workflow)

   • git clone <repository-url> # if you have a repository URL
   • cd basecamp-mcp-server
   • npm install
   • npm run build

2. Run the setup flow to initialize the server and obtain OAuth tokens

   • npx jhliberty/basecamp-mcp-server setup
   • npx jhliberty/basecamp-mcp-server auth

3. Configure OAuth credentials in .env (example values shown; replace with real credentials)

   • BASECAMP_CLIENT_ID=your_client_id_here
   • BASECAMP_CLIENT_SECRET=your_client_secret_here
   • BASECAMP_ACCOUNT_ID=your_account_id_here
   • USER_AGENT="Your App Name (your@email.com)"

4. Generate MCP configurations for your editor/IDE (Claude Desktop or Cursor)

   • npm run config:cursor # For Cursor IDE
   • npm run config:claude # For Claude Desktop

5. Start or run the server as needed (via NPX or a deployment workflow). After initial setup, ensure your AI assistant is restarted so it can load the new MCP configuration.

6. Verify integration in your assistant:

   • In Cursor: Settings → MCP → basecamp should appear with a green checkmark
   • In Claude Desktop: Look for the tools icon in chat and ensure basecamp tools are listed

Tips and issues:

• Ensure the OAuth flow completes successfully and that oauth_tokens.json (or equivalent token storage) exists in your environment.
• Tool visibility may require restarting Claude Desktop or Cursor after configuration changes.
• Use absolute paths for script and config references when wiring MCP into your environment.
• If tools don’t appear, double-check the generated configuration file syntax and the environment variables in your .env file.
• The available tools cover a broad range of Basecamp operations; consult the tool list in the README for exact command names and purposes.
• When using NPX, the package name to invoke may vary depending on the author; ensure you’re referencing the correct package name shown in the setup commands.