For developers and AI agents

Use TurboHog from your AI assistant

TurboHog speaks the Model Context Protocol (MCP), so Claude, Cursor and other AI assistants can read your tasks, projects and schedules, and mark tasks complete when you finish them. Everything they see comes from your own TurboHog account.

What your assistant can do

Tool	Access	What it does
list_upcoming_tasks	Read	Tasks scheduled to start now or later, in calendar order.
list_tasks	Read	Tasks filtered by status, project, schedule or due date, a page at a time.
get_task	Read	One task's details, schedule and dependencies.
list_projects	Read	Your projects, including team projects you can see.
list_schedules	Read	Your schedules and their working hours.
complete_task	Write	Marks a task done, and TurboHog re-plans the rest of your day.

Set it up

The @turbohogapp/mcp-server package runs on your computer and signs in with an API key. It works with any assistant that can start a local MCP server.

Create an API key in TurboHog under Settings, Developers. Give it the tasks, projects and schedules read scopes, and tasks write if your assistant should mark tasks complete. The key is shown once, so copy it.
Add the server to your assistant with one of the setups below.
Restart the assistant. The TurboHog tools appear in its tool list.

Claude Desktop and Cursor

Add this to Claude Desktop's claude_desktop_config.json, or to Cursor's MCP settings:

{
  "mcpServers": {
    "turbohog": {
      "command": "npx",
      "args": ["-y", "@turbohogapp/mcp-server@0.2.1"],
      "env": {
        "TURBOHOG_API_KEY": "thsk_live_..."
      }
    }
  }
}

Claude Code

claude mcp add turbohog \
  --env TURBOHOG_API_KEY=thsk_live_... \
  -- npx -y @turbohogapp/mcp-server@0.2.1

Keep the key secret. If it leaks, revoke it in Settings and create a new one. The package's source and options are on npm.

Things to ask

What are my next three tasks in the Work project?
Which of my tasks are due this week?
What's on my schedule for tomorrow afternoon?
I finished the quarterly report. Mark it done.

Your assistant looks up project and schedule names first, then asks for the tasks that match.

Build your own integration

The MCP tools are built on TurboHog's REST API, which also creates and edits tasks, series and events, and sends webhooks. Read the API reference.