Skip to main content

Stunt Double documentation

Stunt Double is an AI-powered assistant platform. You create actors, AI assistants that can drive a real browser, read email, and work through your product the way a person would, then point them at repeatable jobs: verification checklists, structured user interviews, and multi-step workflows.

This documentation covers the concepts behind the platform and how to operate it, both from the dashboard and programmatically over the API.

Where to start

If you are new to Stunt Double, read these in order:

  • Quickstart walks you from an empty workspace to your first actor and a checklist run.
  • Core concepts explains actors, checklists, interviews, workflows, and insights, and how they relate.
  • API overview and the MCP server show how to automate Stunt Double from your own tools, or from Claude.

How the pieces fit together

A workspace is the top-level container for your team. Inside a workspace you create projects, and inside a project you create actors. An actor has instructions and knowledge, and a set of capabilities (such as browsing or email) that determine what it can do.

Once you have an actor, you give it work:

  • A checklist is an automated verification workflow. The actor runs through a series of steps in a hosted browser and reports what passed or failed.
  • An interview is a structured conversation the platform runs with your users, producing a report you can act on.
  • A workflow chains steps together into a repeatable automation.

Everything an actor does is recorded, so you can replay a session, inspect each step, and feed the results back into your product.

Getting help

If something here is unclear or missing, reach out from the support page. Documentation is versioned alongside the product, so it tracks the current behaviour of the app.