Trello is an MCP server that integrates Trello workspaces directly with AI hosts and model context protocol clients such as Claude Desktop and Cursor. It connects to the official Trello API, allowing software engineers, project managers, and productivity-focused users to manage their kanban workflows using natural language prompts. The server enables AI assistants to inspect boards, create and archive lists, and manage cards across multiple columns. Beyond basic card manipulation, it supports complete checklist workflows, allowing assistants to add, inspect, modify, and delete individual checklist items within tasks. By providing comprehensive read access to boards and full read-write access to lists, cards, and checklists, the server eliminates the need to switch windows when organizing tasks, tracking sprint progress, or maintaining project documentation. It supports local installation via uv as well as standalone network deployment using Server-Sent Events or Docker Compose.
Category: Project & Task Management
Tags: kanban, productivity, task-tracking, trello
git clone https://github.com/m0xai/trello-mcp-server.git && cd trello-mcp-server 3. Copy .env.example to .env and fill in your credentials: bash TRELLO_API_KEY=your_api_key_here TRELLO_TOKEN=your_token_here USE_CLAUDE_APP=true 4. Install dependencies and configure the server for Claude Desktop using uv: uv run mcp install main.py 5. Restart Claude Desktop. 6. For Cursor or SSE clients, set USE_CLAUDE_APP=false in .env, run python main.py, and add the SSE URL http://localhost:8000/sse to your client configuration.Part of MCP Servers
Clone the repository from GitHub, configure your Trello API key and token in a dot-env file, and install the server using the uv package manager command uv run mcp install main.py. Once the installation completes, restart Claude Desktop to load the new tools.
The server provides read access to your Trello boards and full read, create, update, and delete access for lists, cards, checklists, and individual checklist items. Users can ask their AI assistant to search boards, populate cards, and organize tasks.
It integrates directly with Claude for Desktop using standard MCP installation. It also includes an SSE mode running on port 8000 that works with Cursor and other MCP-compatible clients via HTTP streaming endpoints.
No, the server currently supports read-only operations for boards, meaning it can view board summaries and inspect details. Creating, updating, and deleting operations are supported at the list, card, and checklist levels.