What is this website?
Testsy is an AI quality-check service for web and mobile apps. You describe in plain language what should happen and what must be true at the end; Testsy runs that flow in a real browser or on a chosen iPhone, iPad, or Android environment, then returns a pass/fail verdict, step-level evidence, replay media, and a fix prompt you can paste back to a coding agent. It fits small teams shipping with Claude Code, Cursor, or Codex who lack time to maintain brittle UI scripts.
Core features
- Plain-language tests: describe steps and acceptance checks without writing selectors by hand.
- Multi-platform runs: exercise Chrome websites and optional iOS / Android devices and OS versions.
- Evidence and replay: every run includes step replay, values read on screen, screenshots/video, and a copyable fix prompt on failure.
- Coding-agent loop: connect via MCP to Claude Code, Cursor, Codex, and similar clients so an agent can start tests and read results.
- Schedules and CI: save tests to rerun, email on failure, or drive the same suite from CLI, REST, and GitHub Actions.
- Dedicated inbox and phone: signup flows can use a run-specific mailbox; paid plans can cover SMS and call verification scenarios.
Common use cases
- After an AI website builder publishes a preview URL, verify signup or checkout with one written flow.
- After an agent changes one feature, rerun a saved suite to catch regressions in coupons, title persistence, and other paths.
- Mobile teams upload a build, pick device and OS version, and check SMS codes or orientation-sensitive flows.
- Small teams trigger the same checks from CI after each deploy and attach failed replays to tickets.
Who is it for?
- Indie builders who ship with coding agents and want a second opinion on UI behavior.
- Agencies and small teams covering web plus apps without maintaining large brittle selector suites.
- Workflows that feed failure evidence straight back into an agent for repair.
- A poor fit for large QA orgs with mature code-centric E2E stacks that refuse external run quotas.
- A poor fit if you need fully offline, local-only execution with no cloud device environment.
How does it compare to similar tools?
Decide with three questions: whether pay-per-run beats seat-based pricing; whether you need readable expected-vs-seen evidence and shareable replay instead of a green/red light only; and whether an MCP/agent loop is a first-class requirement. If your priority is authoring and versioning Playwright or Appium scripts, a self-hosted framework may be more controllable. If your priority is letting an agent verify the product it just built in plain language, Testsy aligns more closely.
FAQ
Q: What counts as a run on Testsy?
A: Pricing copy defines one run as one test executed on one platform. Running the same test on web and Android counts as two. Blocked runs caused by the site stopping the tester are not billed.
Q: Is there a free allowance?
A: There is a one-time free trial of 10 runs with no card required, then monthly plans that include different run volumes.
Q: Do I need to write code?
A: Not required. You can write plain-language flows and review drafted steps; developers can still drive everything through MCP, CLI, API, or CI.


























