Skip to main content
Glama

CerebroMCP

MCP server for working with the Cerebro project management system. Lets Claude and other LLM clients read data from Cerebro: tasks, messages, users, attachments, and more.

Features

  • List all projects and navigate the task tree

  • Find a task by full path (/Project/Folder/Task)

  • Read task forum messages (briefs, reviews, reports, notes, status changes)

  • View tasks assigned to a specific user

  • Get task attachments or attachments for a single message

  • Read custom tags, checklists, and task dependencies

  • Decode statuses and activity types (reference data)

  • Get the currently authenticated user

For the full list of tools, see AGENTS.md.

Installation

1. Copy the pycerebro library

The library is not included in this repository. Copy the pycerebro folder from your Cerebro installation into the project root:

C:\Program Files\Cerebro\py-site-packages\pycerebro  →  <project folder>\pycerebro

2. Create a virtual environment and install dependencies

cd <project folder>
python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt

3. Verify the installation

python -m py_compile server.py
python -c "import pycerebro; print('OK')"

4. Connect to Claude Desktop

Open the configuration file:

%APPDATA%\Claude\claude_desktop_config.json

Add to the mcpServers section:

{
  "mcpServers": {
    "cerebro": {
      "command": "C:\\path\\to\\project\\.venv\\Scripts\\python.exe",
      "args": [
        "C:\\path\\to\\project\\server.py"
      ],
      "env": {
        "CEREBRO_USER": "your_login@example.com",
        "CEREBRO_PASSWORD": "your_password"
      }
    }
  }
}

Replace C:\\path\\to\\project with the actual project folder path, and set CEREBRO_USER and CEREBRO_PASSWORD to your Cerebro credentials.

Restart Claude Desktop after saving.