A free weekday newsletter for people building with AI, plus fixed-price help when your project stops working. Auth that breaks after publish, Stripe that doesn't unlock, keys in the wrong place: we fix the scoped problem and tell you exactly what we did.
Free newsletter + pre-launch checklist. Unsubscribe any time. Privacy (draft)
The prototype worked in preview. Then you hit the boss fight: going live. These are the problems we see people get stuck on most.
Every prompt burns credits and the app gets no closer to done. We pin down the one bug that keeps coming back and fix it as a single, scoped job. If the loop really needs a rewrite, we'll say so and point you elsewhere rather than quote one.
Site URL still set to localhost, OAuth callbacks that don't match your domain, confirmation emails that never arrive. We fix the redirect and provider settings on your real URL.
Checklist §1
Missing production env vars, hardcoded localhost, CORS errors, silent server failures. We find what differs between your machine and the deployed app and fix it.
Checklist §2
Webhooks that never arrive, signatures that fail on the raw body, test and live secrets mixed up, access granted from a success page anyone can visit. We wire up a verified webhook and test it in Stripe test mode.
Checklist §3
Tables with RLS off, policies that let everyone read everything, grants that don't match what you intended. We review and fix the common failure points and give you a written report.
Checklist §4
Secret keys shipped in client code or git history, AI endpoints with no auth, rate limit, or spend cap. We check for them, fix what's in scope, and tell you what to rotate.
Checklist §5
If your problem doesn't fit a tier, we'll say so and either quote it separately or point you to someone who can help. We never silently expand the scope.
The first 10 jobs get the lower price shown on each card, with the regular price struck through. In exchange, we ask for an honest short review after delivery. Publishing it is optional, and we ask your permission first.
One reproducible problem, in one repo, in one environment.
Get the app live and working end to end, in your own accounts.
Up to two integrations repaired, plus a written review of the usual failure points.
Before you pay, we write down the specific problem we'll fix. If we can't fix that problem, we refund the fee for that job.
Plain-language summary. The terms are still in draft and under review; the final wording goes into your written scope before you pay.
BuildUnstuck
service_role / secret keys, or live Stripe secret keys with us.Prices are USD. The fix-or-refund terms are written into your scope before you pay.
Fill in the short intake form: your repo and host, what happens, what you expected, and what you've already tried. It takes a few minutes.
You get a fixed price and a written scope that says what's in and what's out. If it doesn't fit, we tell you before any work starts. You approve it before paying.
A pull request or patch you apply yourself, plus a short report covering what was wrong, what changed, how to verify it, and what we didn't touch.
Yes. The newsletter and the pre-launch checklist are free. Paid fixes are separate and optional; each issue has one soft call to action, in the same place.
A short intake form: a link to your repo (or a zip) and where it's hosted, which AI tools built it, what happens versus what you expected, any error text or screenshots, what you've already tried, and which third-party services are involved. A staging or preview URL and a test account (with no real customer data) help a lot.
No, and please don't. We work from a staging copy with scoped, revocable tokens. Don't share production credentials, Supabase service_role or secret keys, or live Stripe secret keys. If a key has ever been pasted into a chat, committed to git, or shipped to the browser, we'll recommend rotating it.
No. It's a written review of the most common vibe-coding failure points plus fixes for what's in scope. It is not a penetration test or compliance audit, and it can't promise your app is secure or breach-proof.
Our targets are 2 business days for a single fix, 3–5 for deploy + auth + payments, and 5–7 for the integration and security pass, counted from complete intake (and access, where needed). They're targets, not guarantees.
If we can't fix the specific problem written in the agreed scope, you get a refund of that job's fee. The promise is narrow on purpose: it covers only the scoped problem. It isn't a promise that your whole app will work, and it doesn't cover new bugs or feature requests. It also needs the access and information we ask for. Each tier includes one revision round, and if a problem doesn't fit, we'll tell you up front. The terms are still in draft; the final wording goes into your scope before you pay.
The first 10 jobs get a lower price: $109 instead of $149, $299 instead of $399, or $599 instead of $799. In exchange we ask for an honest short review after delivery. Publishing it is optional and we ask your permission first. After the first 10 jobs the regular prices apply.
Teams that need a full rebuild, apps in regulated areas (health, finance, children's data) that need compliance work, and anyone who needs a guarantee of security. We'll say so and point you elsewhere where we can.
AI tools are part of how we work, which is also why every fix is tested and anything touching auth, payments, or security gets a human review before it reaches you. We tell you what we tested and what we didn't.
Grab the free checklist and run it yourself, or tell us what's broken and we'll scope the fix.