TestSprite

Your coding agent can run all night. It still can't tell if what it built actually works. Today we're open-sourcing the TestSprite CLl (Apache-2.0) A tool your agent calls on its own to test your app end-to-end like a real user, fix what broke, and re-check everything it ever got right. It's the same engine 100,000+ teams already use. github.com/TestSprite/tes… We proved it in public, on a public leaderboard: Most correct app on the board:89% Built by the cheapest model in the field At half the cost of the priciest one You no longer need the biggest, most expensive model to ship software you can trust. Setup is 2 commands: npm install -g @testsprite/testsprite-cli testsprite init That's the last command you'll ever type - from there, your agent runs the tests itself.

Season 03 is officially closed. 🏁 Submissions for the hackathon wrapped up today. The $5,000 prize pool is locked in, and the race for the winning builds is officially on. The final tests have been run, the last commits are shipped, and the loop is complete. Massive thank you to everyone who built with us!

@ayush17_08 @OpenAI Cooking absolute heat!

Built Captrix AI using @Test_Sprite @OpenAI Turn short clips into editable AI captions, fine-tune timing, switch creator styles, and export captioned videos from one browser studio. Live: captrix-ai.vercel.app Code: github.com/ayush002jha/ca… #BuildTheLoop #AI #HackathonS3

@debjit_756 @TestSpriteAI Catching ambiguities early. Brilliant build!

🚀 Submitted **SpecGuard AI** for @TestSpriteAI Build the Loop! 🛡️ AI-powered PRD review workspace that catches ambiguous requirements before development. 🌐 spec-guard-ai-one.vercel.app 💻 github.com/Debjitds/SpecG… #BuildTheLoop #TestSprite

@sahilrakhaiya Permission to ship granted! Incredible work. 🫡

Me: "Ship it." ProofSmith: ❌ "Not verified." Me: "Now?" ❌ "Still not." TestSprite enters the chat. ✅ "Okay... now you can ship it." github.com/SahilRakhaiya0…

@mawa22Q An AI lie detector? Brilliant concept!

Building a Mantiz AI lie detector with the @Test_Sprite CLI. The CLI catches all failed tests, simplifying E2E testing + backend. #TestSprite #Hackathon #DevTools

@36_bagus Proud to power this!

ChurnSense TestSprite list update: 15/15 PASSED. Zero failed. Zero blocked 🔄 After the visual modernization, AI insights, and CSV export hardening landed, reran the full suite with --no-auto-heal. All clean. Agent loop in action for @Test_Sprite S3 #BuildTheLoop #TestSprite

@dreyethh @yunhaojiao Absolute banger of a build, drey!

AI coding agents now run a loop: write a failing test,implement ,rerun,bank the green,commit and log it all in Loop .md But today, a 𝑳𝑶𝑶𝑷.𝒎𝒅 is just pride until something verifies it. An honest agent and a fabricated one can produce identical-looking logs. To trust either, you’d have to manually check every claimed commit against Git and every test run against the checker.That trust is the flaw The rule I set was simple and non-negotiable : a log must prove itself. Every line that claims a commit or a run must be verified against reality and whether a pass actually held. This is why I built 𝐒𝐈𝐆𝐈𝐋: a verification layer for AI coding loops where every line carries its proof 𝐒𝐢𝐠𝐢𝐥 𝐏𝐫𝐨𝐝𝐮𝐜𝐭: sigil-umber.vercel.app 𝐆𝐢𝐭𝐇𝐮𝐛: github.com/Enoch208/Sigil Here is how sigil works 🧵 @Test_Sprite @yunhaojiao

Submit in time: testsprite.com/hackathon-s3

24 hours remaining. ⏱️ Submissions for Season 03 close tomorrow, July 10. The $5,000 prize pool is right there for the taking, and there are still two ways to win. Stop tweaking, run that final test, and ship it. Keep the loop running.

No context window holds every requirement across a long build. An agent ships a feature early, forgets it exists by the next phase, and breaks it without noticing. That's where most regressions come from. The fix isn't more context. It's a suite that keeps checking everything the agent ever shipped. That's the loop's real memory.

@ArdinPrayoga 37/37 is beautiful to see!

Built Timeora the right way for @Test_Sprite Hackathon S3: agent writes → TestSprite CLI verifies on the real app (Vercel + Railway) → we fix → repeat. Just checked the full test list. E2E + backend + new AI assistant flows passing after real failures were caught and resolved

@resquedzn05 Love a guilt-free green build!🌱

Nora is a soft place to study. This week I wired @Test_Sprite's CLI into the project. It caught a test failing on a streak counter — the exact guilt mechanic Nora was built to reject. Somewhere along the way, I'd left it in. 20 tests. Green. 🌱 Coming Soon...

How to join: testsprite.com/hackathon-s3

Season 03 update: the new deadline is July 10. We took a 3 day break over July 4th and were slower to respond right in the middle of your build week. That's on us, so those 3 days go back on the clock. Same rules, same $5,000 prize pool, two ways to win. Keep the loop running.

we crossed 2,500 followers today. The honest update is we're not pausing to celebrate. There's a lot more to ship before 10,000, and we're glad you'll be around for it.

Back in CoderCup, we had four coding agents predict the whole World Cup bracket. Most of them didn't see this one coming: only Kimi and Codex called Portugal vs Croatia for the Round of 32. And they agree on the outcome. Both have it Portugal 2-1, Ronaldo's side into the last 16. The match is on now. Two agents are on the record. See if agents are correct! codercup.ai

@vidamrr Let’s go! Thanks for the support! Time to build! 💻🔥 Good luck to everyone entering!

Devs! No se pierdan el hackathon de TestSprite! He probado el servicio las últimas semanas y es una buena opción para crear pruebas de escenarios. Participen para ganar premios!

@UNgethe Great stuff! Looking forward to see the final app!

good start @Test_Sprite helping fix a few bugs in cognivern.vercel.app

AI coding agents love to say "done." The problem: "done" usually means the code passed its own tests, on the agent's own machine, against mocks. That tells you nothing about whether the deployed app actually works for a real user. Real verification runs against the live app, not mocks. It's the difference between an agent that looks finished and one that is. It's what we measure in CoderCup. Go check it out: codercup.ai

@w3sqr Just found you in our community! 😉

@Test_Sprite I’m in

24 hours left until the loop begins. Building with Claude Code, Codex, or Antigravity? The clock is ticking. Season 3 starts tomorrow. Your agent writes the code; our open-source CLI checks it. Write → verify → fix → verify. The Stakes: >$5,000 total pool up for grabs. >$3,000 for the best projects over 7 days. >$2,000 standing bounty for open-source CLI improvements ($100+ per merged PR). Build starts June 30. Get your agents ready.