Prompt → test
"Log in and add a blue mug to the cart." becomes a real, readable Playwright script — not a brittle recording.
Describe what a user does. Testy turns it into a Playwright test, runs it from your CI, and fixes it when your frontend team ships on Friday.
{PASSWORD}No jargon, no dry test-framework walls of text. Just what breaks and what doesn't.
"Log in and add a blue mug to the cart." becomes a real, readable Playwright script — not a brittle recording.
Flaky gets fixed. Broken apps stay broken. We're paranoid about not fixing tests that should fail.
Write {PASSWORD} once. Testy resolves it per environment — the model never sees the value.
One curl in your pipeline. NDJSON streams results as they land, then a clean pass/fail summary.
Toggle it on, tell us who logs in, and every test starts from a signed-in session.
See every run, every screenshot, every trace — and a score for which tests are getting shaky before they break.
When a prompt is ambiguous, the agent asks you a question instead of guessing.
Every generated or hand-edited script is a version. Diff it, restore it, audit exactly what the healer changed.
Three steps, no YAML required.
Write what a user does, in a sentence or two.
An agent explores your site, writes the script, and runs it until it passes.
From the dashboard, on a schedule, or with one curl in CI.
Start free. Upgrade when you outgrow it.
Have a gift code? Redeem it after signing in.
Refunds, no questions asked.
Extremely real feedback from extremely real people.
"I used to write Playwright selectors by hand. I now write poems."
— A recovering QA engineer(not a real person)"It fixed a test at 3am and didn't even wake me up. Rude, honestly."
— CTO, a startup that exists(not a real person)"Our flaky tests are now just… tests."
— Someone who stopped rerunning CI(not a real person)"Testy asked me a clarifying question. My coworkers never do."
— Senior developer, 14 open PRs(not a real person)"10/10, would {PASSWORD} again."
"Our onboarding doc now just says "ask Testy"."
— Engineering manager, mildly concerned(not a real person)Fewer 2am pages about a login button nobody moved on purpose.
New teammates ship a smoke test in minutes, not a sprint.
Cheaper than a QA contractor's coffee budget.
You describe a user journey in plain language. An agent explores your site, writes a Playwright script, and verifies it actually passes before handing it back to you.
A target is one of your apps (with a base URL and secrets). Environments are variants of it — staging, production — that can override individual secrets.
Secrets are encrypted at rest with AES-256-GCM. The model only ever sees secret names, never values — substitution and redaction both happen inside the runner, never in a prompt or a log.
Any manual or CI-triggered run. Retries the healing agent does on its own to fix a test don’t count against your quota.
Flaky means it sometimes passes, sometimes fails on the same code — we make it deterministic. Broken means it fails consistently — we only fix the test if the scenario still works by hand; otherwise we leave it failing and tell you why.
Yes. Every test is a real, readable Playwright script you can open, edit, and save as a new version.
Start on Free. Move up when you outgrow 3 targets, 30 tests, or 300 runs a month.
No questions asked. Email us and we’ll refund the current period.
Chromium today. Firefox and WebKit are on the roadmap.
One curl, with your API token and target/environment in the query string. We stream results as NDJSON so your pipeline gets a clean pass/fail.