Experimental MCP Server for CommCare Connect (ccc)

What is Experimental MCP Server for CommCare Connect (ccc)?

This experimental MCP server integrates with CommCare Connect's global stats API, allowing Claude Code and other MCP clients to query program dashboard data. It is built on the Python MCP SDK and is intended for developers and administrators who need programmatic access to earned amounts, paid amounts, active users, and visit statistics.

How to use Experimental MCP Server for CommCare Connect (ccc)?

To use it with Claude Code, run claude in the server folder. Modify the .mcp.json declaration for other tools. You need uv installed. Obtain an authentication token via the superuser admin page or a curl command with your username/password.

Key features of Experimental MCP Server for CommCare Connect (ccc)

• Queries CommCare Connect’s global stats API
• Filters by date ranges and program/organization IDs
• Returns earned/paid amounts, active users, visit stats
• Works with Claude Code and configurable for other MCP tools
• Supports local development and staging environments

Use cases of Experimental MCP Server for CommCare Connect (ccc)

• Monitor earned and paid amounts across programs and organizations
• Track active user counts over custom date ranges
• Retrieve visit statistics for program dashboards
• Automate program reporting with natural language queries via Claude

FAQ from Experimental MCP Server for CommCare Connect (ccc)

What can this MCP server query?

It queries the global stats API used by the program dashboard, returning amounts earned/paid, active users, and visit statistics. You can optionally filter by date ranges and specific program/organization IDs.

How do I get an authentication token?

As a superuser, visit https://connect.dimagi.com/admin/authtoken/tokenproxy/ to find a token associated with your email. Alternatively, run curl -X POST -d "[email protected]&password=***" https://connect.dimagi.com/auth-token/ to create one.

What dependencies are required?

You need uv installed on your system. The server itself uses the Python MCP SDK.

Can I test against a development or staging server?

Yes. Change the SERVER_ENDPOINT in .mcp.json to http://localhost:8000 or https://connect-staging.dimagi.com/ and obtain a token from that environment.

Is this production‑ready?

No, it is explicitly experimental and built on the Python MCP SDK. Current functionality is limited to querying the global stats API.