This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
To-do boards for AI agents, served as an MCP server. An agent lays out a plan, moves each to-do through its status, records why one is blocked, and asks what to do next. A person sees the same boards in a board UI that renders inside MCP Apps hosts.
It runs on your own machine: one SQLite file, no account, no cloud service.
Run it with uv (Python 3.13 or later):
uvx todosforagentsThat serves MCP over stdio. To add it to Claude Code:
claude mcp add todosforagents -- uvx todosforagentsFor Claude Desktop, the MCP bundle @nimblebraininc/todosforagents is available from
mpak. It runs the same package and asks where to keep the database.
Any other MCP client that starts a stdio server can run uvx todosforagents the same way.
To serve HTTP on 127.0.0.1 instead, run uvx todosforagents --http; HTTP requires a
bearer token, which uvx todosforagents --print-token prints. See
docs/local.md.
Related MCP server: crunchy-work
Everything is in one SQLite file, todos.db, in your user data directory:
OS | Default location |
macOS |
|
Linux |
|
Windows |
|
Set TODOS_LOCAL_DB to use another file. Keep it on a local disk, not a network share.
Several clients may use the same file at once. Upgrades, backups and the HTTP token are
covered in docs/local.md.
Tool | What it does |
| Add one to-do (to the default board unless one is named) |
| Add several to-dos in one call, optionally as the steps of a goal |
| Change a to-do's title, details, priority, labels, assignee, due date or goal |
| Mark a to-do in progress |
| Mark a to-do done |
| Mark a to-do blocked, with the reason |
| Put a to-do back to open |
| Mark a to-do as won't do |
| Set the order of one status lane on a board |
| Hide a to-do from default views, status unchanged |
| List or search to-dos, with filters and cursor paging |
| Read one to-do whole, with its steps and recent activity |
| The highest-priority open to-do: what to work on next |
| A board's lane counts and what needs attention |
| List the boards |
| Create a board |
| Change a board's name, description or lanes |
A to-do's status (open, in_progress, blocked, done, cancelled) changes only through
the five status verbs. Every result carries a next_step telling the agent what to do with
it, and the server also serves a skill (skill://todos/SKILL.md) with the rules for choosing
a tool.
Every board has a short key, and every to-do a key made of its board's key and its number
on that board: MAG-3 is the third to-do on board MAG. The default board's key is TODO.
Any argument that takes a to-do id takes its key, and any argument that takes a board id
takes the board's key, so an agent can pass on "MAG-3" exactly as a person said it.
get_board_summary and get_todo are bound to the board UI (ui://todos/main). A host that
supports MCP Apps shows the board, or the to-do, when the
agent calls them. The board reads and changes data only through the same tools.
docs/architecture.md: the core, runtimes, data model, the UI
binding, errors and annotations.
docs/local.md: the database file, several processes, the HTTP token,
upgrades and backups.
CONTRIBUTING.md: building and testing.
MIT. See LICENSE.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Task & board management for AI agents + humans. Kanban, comments, digests via MCP.
Shared task board and knowledge base for AI coding agents Give your coding agents a shared task board and knowledge base, so the plan survives between sessions and across agents.
A task list shared by you and your coding agents (Claude Code, Cursor, Codex and any MCP client). Agents plan each task into steps, report live progress to a board, and block on questions you answer from your phone.
Persistent memory, knowledge graph, task management and more for AI agents via MCP.
