AI Agent Integration
Give your AI agents a kanban board
Kanboard ships a remote MCP server. Claude Code, Codex, Cursor, and any MCP-compatible agent can read your board, create tasks, and move work across columns — no local install, one command to connect.
1. Create an API key
Go to Settings → Integrations in Kanboard and create an API key. Keys require a Pro subscription. Copy the key — it's shown once.
2. Connect your agent
Claude Code
claude mcp add --transport http kanboard https://kanboard.io/api/mcp \
--header "Authorization: Bearer YOUR_API_KEY"Codex
# ~/.codex/config.toml
[mcp_servers.kanboard]
url = "https://kanboard.io/api/mcp"
http_headers = { "Authorization" = "Bearer YOUR_API_KEY" }Cursor / other MCP clients
{
"mcpServers": {
"kanboard": {
"url": "https://kanboard.io/api/mcp",
"headers": { "Authorization": "Bearer YOUR_API_KEY" }
}
}
}Set up your client
Step-by-step guides with verified config snippets for every major MCP client:
One CLI command
mcp-remote proxy in JSON config
mcp.json entry
codex mcp add with a bearer-token env var
mcp_config.json entry (note: serverUrl)
.vscode/mcp.json with an encrypted input
context_servers in settings.json
settings.json entry (note: httpUrl)
cline_mcp_settings.json entry
opencode.json entry
3. Put your agent to work
Ask your agent things like:
- “Look at my Kanboard board and pick up the top task in To Do.”
- “Break this feature into tasks and add them to my board.”
- “Move everything you finished today to Done.”
Optionally, add this to your repo's AGENTS.md so agents use the board by default:
## Task board
This project tracks work on Kanboard (MCP server: "kanboard").
- Before starting work, check the board with get_board and pick
the top task in "To Do".
- Move the task to "In Progress" while working, "Done" when merged.
- File newly discovered work as tasks in "To Do".Available tools
| get_board | Full snapshot of a project's columns and tasks |
| list_projects | List all your projects |
| create_project | Create a project with default columns |
| list_tasks | List tasks, optionally by project |
| get_task | Read a task with its description |
| create_task | Add a task to any column |
| update_task | Edit a task's title or description |
| move_task | Move a task across columns |
| delete_task | Remove a task |
Prefer raw HTTP? The same operations are available on the REST API.
FAQ
What is the Kanboard MCP server?
It's a remote Model Context Protocol server hosted at https://kanboard.io/api/mcp. Any MCP-compatible AI agent — Claude Code, Codex, Cursor, and others — can connect to it with an API key and read your board, create tasks, update them, and move them between columns.
Do I need to install anything?
No. The MCP server is remote (Streamable HTTP), so there's nothing to run locally. You add it to your agent with one command and authenticate with a Kanboard API key.
Which AI agents work with Kanboard?
Any client that supports remote MCP servers over Streamable HTTP: Claude Code, Claude Desktop, Codex, Cursor, Windsurf, and more. If your tool speaks MCP, it works with Kanboard.
Is the MCP server included in the subscription?
Yes. The MCP server and REST API are included in Pro, Kanboard's single plan at $12/month (or $96/year) with unlimited projects and tasks.
What can agents actually do on my board?
The server exposes tools to list projects, snapshot a full board, create projects, and create, read, update, move, and delete tasks. A typical flow: you plan tasks in To Do, your agent picks one up, moves it to In Progress, does the work, and moves it to Done.