Connect your agent

The gcontext connector is a Python package on PyPI. It runs locally on your machine and bridges your AI client to your cloud workspace.

Prerequisites

• Python 3.10+ (check with python --version)
• An AI client that supports MCP: Claude Code, Claude Desktop, Cursor, Codex, or any MCP-compatible tool
• A gcontext account at app.gcontext.ai

You do not need to install the connector manually. The uvx command (included with Python's uv tool) runs it directly from PyPI.

Step 1: Get your token

1. Open your dashboard at app.gcontext.ai.
2. Go to Setup > Connection.
3. Click Mint token, give it a label (e.g. "laptop"), and copy the token. You will only see it once.

Step 2: Add the MCP server to your client

Pick your client and paste the config:

Claude Code

Run this in your terminal:

claude mcp add-json gcontext '{"type":"stdio","command":"uvx","args":["gcontext-mcp"],"env":{"GCONTEXT_TOKEN":"your-token-here"}}'

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "gcontext": {
      "command": "uvx",
      "args": ["gcontext-mcp"],
      "env": {
        "GCONTEXT_TOKEN": "your-token-here"
      }
    }
  }
}

Cursor

Add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "gcontext": {
      "command": "uvx",
      "args": ["gcontext-mcp"],
      "env": {
        "GCONTEXT_TOKEN": "your-token-here"
      }
    }
  }
}

Codex

Add to ~/.codex/config.toml:

[mcp_servers.gcontext]
command = "uvx"
args = ["gcontext-mcp"]
env = { GCONTEXT_TOKEN = "your-token-here" }

Step 3: Verify

Open your AI client and ask your agent something like:

Using gcontext, show me an overview of my workspace.

The agent should connect and return your workspace snapshot. On the dashboard, the connection pill in the sidebar turns green.

If it does not connect, check:

• Your token is correct and not expired
• uvx is available (install with pip install uv or brew install uv)
• Your client is configured to use MCP servers

Next: Your first workspace.