Skip to main content

Integration guides

Stunt Double's core offering is simple: AI user personas (actors) that drive real browsers through real products and report back with evidence. Anything with a reachable URL can be tested, so Stunt Double is not a destination you visit but a validation layer that plugs into wherever the work already happens.

These guides show how to wire it into the most common workflows so testing, research, and standards enforcement happen automatically, at the moment a design or change exists, rather than as a separate chore afterwards.

The guides

  1. Claude design Design an artifact in Claude, publish it, and validate it with personas in the same conversation.
  2. Figma Test prototypes with personas and re-check flows automatically when files change or the library is published.
  3. Figma Make Publish a Make site and interview personas against it before an engineer touches it.
  4. Codebases Verify changes on preview deployments from Claude Code, Cursor, or GitHub, and report results on the pull request.
  5. Validation Turn brand, tone-of-voice, legal, continuity, and design-system standards into checks that run on every deploy and design update, with ready-to-use actor and workflow examples.

The integration model

Three mechanisms make Stunt Double feel native in other tools:

  1. The MCP server (https://app.stuntdouble.io/api/mcp). Any MCP-compatible agent (Claude, Claude Code, Cursor, Windsurf, custom agents) gets the full toolset: create and run checklists, launch persona interviews, triage feedback, comment on PRs. The server ships workflow prompts (validate_design, verify_change, run_user_research, triage_feedback, setup_guardrails, and the validation prompts check_brand, check_design_system, check_compliance, check_continuity) that appear as slash commands in most clients. See the MCP server reference for setup.
  2. Connections and event triggers. Workspace connections (GitHub, Figma, Linear, Slack, Vercel) let workflows fire on GitHub, Figma, Vercel, Linear, and Slack events, so checks re-run on pull requests, deploys, and design updates without anyone asking.
  3. The dashboard (app.stuntdouble.io) for anything visual: watching session recordings, reading interview reports, managing actors and connections.

Where each role feels it

  • Designers get usability findings while the design is still cheap to change: persona interviews against prototypes, with transcripts and evidence instead of hunches.
  • Product managers get research at decision speed and feedback triage that clusters and reproduces what users report.
  • Engineers get a user-level regression net: checklists that exercise real flows on preview deployments, results as PR comments, and event-triggered re-runs on deploys.
  • Brand, legal, and design-system owners get enforcement instead of periodic review: standards codified once and checked on every deploy or design update. See the validation guide.