Every PR, tested.
No tests written.
Kind reads your pull request, spins up your app in a sandbox, and tests it end-to-end. Or paste any URL for an instant quick test. PR results land as a comment before you review — quick tests show up in your dashboard.
How it works.
Reads your diff
Kind sees what changed in your PR: the files, the framework, the app structure. It generates test scenarios for the code you actually touched.
Spins up a sandbox
Your repo gets cloned into an isolated container with dependencies installed and secrets injected. Ephemeral, destroyed after every run.
Tests in a real browser
Playwright navigates your app, fills forms, clicks buttons, and verifies the DOM. Screenshots are captured at every step.
Reports on your PR
Pass/fail results, failure details, and screenshots posted as a PR comment. Your team sees the verdict before the code review.
Any stack.
Zero config.
Kind detects your framework, finds your preview deployment, and adapts its tests. No config files, no test specs, no CI changes.
See what Kind sees.
Every test step, screenshot, and failure reason, visible in the dashboard and posted directly on your pull request.
Run detail
Test Plan
Preview loaded at checkout-redesign-a1b2c3.vercel.app
Cart updated, checkout page rendered with saved address
Timed out waiting for confirmation page after clicking submit. Button fires but navigation never completes.
Empty cart correctly blocked with validation message
PR comment
Test Results
Results
| # | Test | Status | Details |
|---|---|---|---|
| 1 | Navigate to preview deployment | ✅ passed | Preview loaded successfully |
| 2 | Add item to cart and proceed to checkout | ✅ passed | Cart updated, checkout rendered |
| 3 | Submit order with saved address | ❌ failed | Navigation never completes after submit |
| 4 | Submit order with empty cart | ✅ passed | Validation message shown |
Issues Found
Submit order with saved address: Clicking the submit button fires the handler but navigation to the confirmation page never completes. Expected redirect to /order/confirmation. Likely a missing await on the checkout mutation.
Summary
3 of 4 tests passed. The checkout submit flow is broken — the order mutation fires but the confirmation redirect never happens. Edge cases and cart validation work correctly.
Two minutes to first test.
Connect GitHub, pick a repo. Your next pull request gets tested automatically.
“Testing can be used to show the presence of bugs, but never to show their absence.”
— E. W. Dijkstra
Kind catches the regressions that slip through code review. Broken flows, missing pages, failed checkouts, caught before they reach production.
Questions?
Everything you need to know before getting started.
Do I need to write any tests?
No. Kind reads your PR diff and generates test scenarios from the code change. No test files, no config, nothing to maintain.
What stacks are supported?
Next.js, Vite, Express, Rails, Django, Flask, Go, and more. Kind auto-detects your framework and adapts its tests.
How does Kind access my code?
Through the GitHub App. Your repo is cloned into an ephemeral sandbox that's destroyed after every run. Secrets are encrypted with AES-256-GCM.
Does it work with preview deployments?
Yes. Kind detects Vercel, Netlify, and Railway preview URLs automatically, waits for the deploy to finish, then tests against the live preview.
Can I trigger a test manually?
Yes. Comment "@kind-agent test" on any PR to start a run. You can also add custom instructions for specific scenarios you want tested.
What is a quick test?
Quick test lets you paste any URL and test it instantly — no PR needed. Add instructions like "sign up with a test email" or "try to checkout with an empty cart" and Kind runs the scenario in a real browser. Results appear in your dashboard with screenshots and a full summary.
What does a test run cost?
Each run uses credits based on complexity. The free tier includes 15 one-time credits. Paid plans start at $25/month with 100 monthly credits.