kwiklytrellomcp
A Model Context Protocol (MCP) server that exposes Trello boards, lists, cards, and checklists to AI assistants like Claude and Cursor.
63 commits over 14 days
What it is
This tool bridges Trello's project management capabilities with AI coding assistants by implementing an MCP server. It allows Claude Desktop, Cursor, and other MCP-compatible clients to read and manipulate Trello boards, lists, cards, and checklists through natural language. The server supports both direct Claude Desktop integration and a standalone SSE (Server-Sent Events) mode, and can be deployed via Docker or to Coolify for remote access.
Why I built it
Built to enable AI assistants to interact with Trello workspaces, likely motivated by the need to automate project management tasks or query Trello data during development workflows. The extensive Coolify deployment documentation suggests this was intended for production use as a self-hosted service accessible from multiple clients.
Tech stack
Python 3.12+ with the MCP SDK, using the uv package manager. Trello REST API integration via custom API utilities. SSE server mode uses uvicorn and starlette. Containerized with Docker and Docker Compose. Deployment-ready for Coolify with health checks, non-root user, and SSL support.
Timeline & effort
Last active Oct 14, 2025 — 63 commits across 14 active days (spanning 195 days).
Developed over 195 days with 63 commits across just 14 active days (mostly in April and May 2025, with sporadic updates through October). This intermittent pattern—averaging ~4-5 commits per active day—suggests focused build sessions interspersed with long breaks, consistent with a side project or utility built and refined as needed.
Did it turn out useful?
The project is fully functional and production-ready, with comprehensive deployment guides, Docker support, health checks, and Coolify configuration. The extensive documentation (deployment guides, checklists, quickstarts) and production features (non-root Docker user, SSL setup, DNS instructions) indicate this was shipped for real use. While there's no evidence of external users or stars (private repo), the polish and deployment infrastructure point to a working tool actively used by the author or a small team.