Skip to content
Vetroscope Docs

Vetroscope MCP

Vetroscope MCP connects AI assistants to your tracked time through the Model Context Protocol. Once added, you can ask Claude Desktop, Claude Code, Cursor, or any MCP-aware client things like “how much After Effects did I do this week?” or “what was I working on Tuesday morning?” — and it answers from your real data.

It reads your local Vetroscope data directly on your machine: read-only, no account, no cloud round-trip, works offline. Product overview on vetroscope.com/mcp.

Setup

Section titled “Setup”

Claude Code:

Terminal window
claude mcp add vetroscope -- npx -y vetroscope-mcp

Claude Desktop — add to claude_desktop_config.json:

{
  "mcpServers": {
    "vetroscope": {
      "command": "npx",
      "args": ["-y", "vetroscope-mcp"]
    }
  }
}

Cursor — the same command/args block in ~/.cursor/mcp.json.

Requires Node.js on the machine. The full tool reference and more client configs (Zed, ChatGPT) are in the Vetroscope MCP repository.

MCP, Local API, or Chat?

Section titled “MCP, Local API, or Chat?”
  • Vetroscope Chat — built into the app, zero setup, Pro.
  • MCP — the same kind of questions, inside AI tools you already use; free with the app.
  • Local API — raw data for code you write yourself.