MCP Kanban Memory
@eyalzh
MCP server providing kanban-based task management memory for complex multi-session workflows with AI agents
Overview
What is MCP Kanban Memory?
MCP Kanban Memory provides kanban-based task management state for AI-driven development. It structures an AI agent’s work by documenting and saving tasks in a kanban board, both during planning and execution sessions. It is similar to other task management MCP tools but adds column capacity limits, an embedded SQLite database, a web UI for observation, and predefined prompts.
How to use MCP Kanban Memory?
Clone the repository and run npm ci and npm run build for the shared/db and mcp-server directories. Add the server to your MCP client (e.g., Claude Desktop) using the provided JSON configuration, setting the MCP_KANBAN_DB_FOLDER_PATH environment variable to a writable folder. Use the predefined MCP prompts (create-kanban-based-project or make-progress-on-a-project) or ask the AI directly to create a board, add tasks, move them, and resume work later. A Docker installation option is also available.
Key features of MCP Kanban Memory
- Column capacity / work-in-progress limits
- Embedded SQLite database for persistence
- Web UI to observe and modify tasks
- Predefined prompts for starting and resuming workflows
Use cases of MCP Kanban Memory
- AI-driven development planning: create a board, divide a project into tasks, and track completion.
- Resuming a paused session: find
