Introduction to the Donobu API

What the Donobu API is, its local-first architecture, and when you might need to use it directly.

What is the Donobu API?

The Donobu API is a local REST API that runs at http://localhost:31000 when Donobu Studio is running. It powers both Donobu Studio and the Donobu Playwright Extension.

When you create a flow in Studio or call page.ai() in a Playwright test, the request goes to this API, which orchestrates the browser automation — launching the browser, running AI-driven actions, capturing screenshots, and recording results.

Local-first architecture

The API runs entirely on your local machine. No data is sent to external servers (other than LLM provider API calls for AI decision-making). This means:

  • Privacy — Your automation data stays on your machine
  • Speed — No network round-trips to a remote server
  • Offline-capable — Works without internet access (except for LLM API calls in autonomous mode)

When to use the API directly

Most users should interact with Donobu through Studio or the Playwright Extension rather than calling the API directly. The API is useful for:

  • Custom integrations — Building your own tooling on top of Donobu
  • Programmatic flow management — Creating, listing, or deleting flows from scripts
  • Advanced configuration — Managing GPT providers, environment variables, and agent assignments programmatically
  • MCP server — The Donobu MCP server uses the API to bridge AI coding assistants with browser testing

If you're just getting started, we recommend beginning with Donobu Studio (visual, no-code) or the Playwright Extension (TypeScript tests).

API basics

  • Base URL: http://localhost:31000
  • Authentication: None required (the API only accepts local connections)
  • Format: JSON request and response bodies
  • Health check: GET /api/ping — returns {"status": "pong"}

See the API Reference for the complete list of endpoints.