kanban

This MCP server provides a kanban-based task management state for AI-driven development. It exposes a set of prompts and an API to create kanban boards, add tasks, move tasks between columns with WIP limits, and query boards or tasks. You can use it to guide the AI to plan work as a board, persist tasks with descriptions written in markdown, and observe progress through a web UI. The included web UI lets you observe boards, adjust tasks manually, and monitor column capacities in real time. To resume work in a follow-up session, locate the relevant board and fetch its current tasks and columns, then continue selecting the next task to work on and move it through the workflow.

Prerequisites:

• Node.js and npm installed on your system
• Access to clone the repository for the Kanban MCP server

Installation steps:

1. Clone the repository: git clone https://github.com/<org>/<repo>.git cd <repo>

2. Install dependencies for the shared DB and the MCP server: npm ci --prefix shared/db npm ci --prefix mcp-server

3. Build the shared DB and the MCP server: npm run build --prefix shared/db npm run build --prefix mcp-server

4. Run or configure the MCP server in your client integration: For Node usage, the server is typically started via the Node command pointing to the built server.js, e.g. in client config: { "mcpServers": { "kanban-mcp": { "command": "node", "args": ["/path/to/repo/mcp-server/dist/server.js"], "env": { "MCP_KANBAN_DB_FOLDER_PATH": "/path/to/db" } } } }

Note: Replace /path/to/repo with the local path to the cloned repo and /path/to/db with a writable folder that will store the database files.

Tips and tips:

• The Kanban server uses a SQLite embedded DB stored in the folder specified by MCP_KANBAN_DB_FOLDER_PATH. Ensure the path exists and is writable.
• When using Docker, you can mount a host directory to /mcp (as shown in the Docker run example) to persist the database.
• Web UI is available after starting the web-server and can be accessed at http://localhost:8221, where you can view boards, tasks, and WIP limits.
• The API supports creating boards, adding tasks with markdown content, moving tasks between columns with optional reasons, and retrieving board/task information for AI-driven planning and progress tracking.
• If you change the environment or paths, restart the MCP server so config changes take effect.