WEEEK MCP with the knowledge base
weeek-mcp
A WEEEK MCP server: tasks via the public API and full knowledge-base CRUD via the internal API with a saved browser session
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- Stores a browser session for access to WEEEK's internal API
- Can create, edit and delete knowledge base documents
Install
Manual install
pip install "weeek-mcp[kb]" && playwright install chromium && weeek-mcp-login && claude mcp add weeek -s user -e WEEEK_API_TOKEN=... -- weeek-mcpInstallation for other MCP clients; [kb] can be dropped if the knowledge base is not needed.
This is third-party code. Review the repository files before installing.
What it does
The server covers two parts of WEEEK with different access modes. Tasks, boards and columns go through the official public REST API: the full task lifecycle, assigning and unassigning members. The knowledge base has no official public API, so the server calls WEEEK's internal JSON API through a saved browser session obtained via a one-time weeek-mcp-login sign-in with Playwright. Knowledge base documents render to Markdown and publish as weeek-kb://<id> MCP resources, so they can be selected as content right in the MCP client rather than only linked. In-place body editing speaks WEEEK's collaborative Hocuspocus/Yjs protocol directly, since the content lives in a shared editing document.
Who it is for. For people who need not just WEEEK tasks but also knowledge base documents right in the agent's context.
Good fit when
- You need access to both WEEEK tasks and the knowledge base from one server
- You want knowledge base documents pulled in as content, not just a link
- You use Claude Desktop and want a single-file .mcpb install
Not a fit when
- You do not want to sign into WEEEK via a Playwright browser session just for the knowledge base
- You only need tasks, a simpler server without Playwright would do
Example request
Find the onboarding document in the knowledge base and create a task to review it for accuracyLimitations
Knowledge base access requires Playwright and a saved session from a one-time sign-in; with 2FA or SSO, sign-in must be repeated manually via weeek-mcp-login. Without an email and password in settings, the server cannot re-authenticate itself when the session expires. Tasks work without Playwright, using only WEEEK_API_TOKEN. Pulling documents through an attachment menu depends on the specific client's MCP Resources support, not universal.
How to disable. Remove the .mcpb extension in Claude Desktop via Settings → Extensions, or remove the weeek-mcp server from another MCP client's configuration.
MCP
- Transport
- stdio
- Authentication
- API key
| Environment variables | |
|---|---|
| WEEEK_API_TOKEN required, secret | A personal WEEEK token for tasks via the public API. |
Security check
- Stores a browser session for access to WEEEK's internal API
- Can create, edit and delete knowledge base documents
README in short
The English README, with a separate Russian version, covers two feature areas: tasks via the public REST API and the knowledge base via WEEEK's internal API with MCP resources. It shows two install paths: .mcpb for Claude Desktop with no Python, and pip for other clients (Claude Code, Cursor, Windsurf, VS Code and others) with a one-time knowledge-base sign-in via weeek-mcp-login and Playwright. It separately explains that pulling documents through attachments depends on client MCP Resources support.
FAQ
Why does the knowledge base not use the official API?
WEEEK has no public API for the knowledge base, so the server uses the internal JSON API through a saved browser session from a one-time sign-in.
Is Python needed for the Claude Desktop install?
Not for tasks: Desktop runs the .mcpb server with its own bundled uv. Python is only needed for a pip install in other clients or for the one-time knowledge-base sign-in.
Related
A self-hosted knowledge base with block-level references and a built-in MCP server for connecting AI agents to your notes
A CLI for every Google Workspace API with JSON output and agent skills: Drive, Gmail, Calendar, Sheets and more
Local search over Markdown notes, docs and meeting transcripts: keywords, semantic search and reranking, with an MCP server
A task manager for AI-driven development: breaks a PRD into dependent tasks and guides the agent through them via MCP or CLI