An AI Testing Agent That Lives in Your Terminal.
Install the TestSprite CLI, point it at a live URL or API, and get real browser and backend test coverage without leaving the command line. Plan tests from a PRD or API spec, run them in CI, and get results back as structured JSON your scripts — or your coding agent — can act on.
One Command Wires It Into Your Coding Agent
Save What You Broke
Run testsprite test rerun <testId> after every change and get a failure bundle back the moment something breaks — before you find out from a user.
Understand What You Want
Scaffold a plan with testsprite test create --plan-template, or point it at your PRD or API spec (OpenAPI, Swagger, Postman) so it can plan coverage on its own.
Validate What You Have
testsprite test create --type frontend --plan-from plan.json --run --wait — real browser tests via Playwright, or dependency-aware backend API tests, run against a live URL.
Suggest What You Need
Every failure ships as one JSON object — failing step, screenshot, DOM snapshot, root-cause hypothesis, fix recommendation — ready for your scripts or your coding agent to act on.
$ npm install -g @testsprite/testsprite-cli
$ testsprite setup
$ testsprite test create --project prj_8f2a --type frontend \
--plan-from checkout-plan.json --run --wait --output json
✓ TC001_Login_Rejects_Empty_Password passed 1.2s
✗ TC002_Checkout_Applies_Promo_Code failed 3.4s
→ failure bundle: testsprite test failure get TC002
$ testsprite test rerun TC002 --wait --output json
✓ TC002_Checkout_Applies_Promo_Code passed 2.9s
Test Anywhere You Can Open a Terminal
Local dev, CI runner, or inside your coding agent's own shell tool — the same testsprite command works everywhere, with --output json for scripting and stable exit codes for gating.
Built for Terminal-First Teams
Batch & Schedule Runs
testsprite test run --all --project <id> re-runs your whole suite in one command — wire it into cron or your CI scheduler.
Test Plans as Code
Version your test plans as JSON and review them in a pull request like any other change — plan-from, plan put, plan-template.
Free to Start
Apache 2.0 licensed CLI — install it and run your first test for free, no seat limits to get going.
Frontend + Backend, One CLI
Real browser tests via Playwright for frontend, dependency-aware API tests for backend — --type frontend or --type backend, same command shape.
Trusted By Businesses Worldwide
"TestSprite offers rich test case generation, clear structure, and easy-to-read code. It also supports simple online debugging with the ability to quickly expand by generating new test cases."
"TestSprite's automation helps us reduce tons of manual work. The developers can easily catch and resolve bugs earlier in the development process."
FAQ
What does the TestSprite CLI actually run?
Real tests against a live URL or API. Frontend tests drive a real browser via Playwright; backend tests send real HTTP requests with dependency management. Nothing runs against mocks — --type frontend or --type backend picks which.
How do I connect it to my coding agent?
testsprite agent install <name> wires it into Claude, Cursor, Copilot, Codex, Windsurf, Cline, Antigravity, or Kiro directly from the CLI — no separate server or protocol to configure.
Can I run it without an AI agent at all?
Yes. testsprite test create --plan-from plan.json --run --wait --output json works from any terminal or CI runner. The CLI doesn't require an agent in the loop — it's just as useful driven by a human or a script.
What happens when a test fails?
You get a failure bundle: one JSON object with the failing step, a screenshot, a DOM snapshot, a root-cause hypothesis, and a fix recommendation, all tagged with a matching snapshot ID — grab it with testsprite test failure get <testId>.
Is it free to try?
Yes. npm install -g @testsprite/testsprite-cli and testsprite setup gets you running in a couple of minutes, and the CLI itself ships under an Apache 2.0 license.