We scope it together
A 20-minute call becomes a signed SCOPE.md: the screens, the data, one core flow, and exactly what "live" means. No ambiguity, no scope creep.
Fixed price · Seven days · Deployed
Your MVP.
Built by an AI agent.
Live in 7 days.
Claude Code builds and deploys a working version of your product in one week, directed and reviewed by a human who's shipped dozens. $3,900, fixed. Not live on day 7? You don't pay.
20 minutes. No pitch. You leave with a written scope either way.
Replay of the build log. Speed-compressed, output unedited.
7days, every time
$3,900fixed, on the page
100% of the repo is yours
0paid if it isn't live on day 7
How it works
A human at both ends. An agent in the middle.
The agent is why it's one week and one price. The human is why it's shippable.
The agent builds it
Writes the code, the migrations, and the tests. Runs them. Fixes what fails. Opens a pull request every day from a starter repo with auth, payments, and deploy already wired.
We review, you approve
Every PR gets a human review before it merges. You get a 2-minute video each evening. You test the preview on day 6. It deploys to your domain on day 7.
Watch a build
Seven days, one log.
What a week actually looks like. Pick a day.
This is for you if
- You have a validated idea: customers, a waitlist, or a pilot
- Your product has one core flow someone would pay for
- You need a demo for investors or first users in weeks, not months
- You tried Lovable or Bolt, got 70% there, and stalled
- You want to own the code and keep building after
This isn't for you if
- You need a native iOS or Android app on day 7
- You handle regulated data (HIPAA, PCI beyond Stripe)
- Your v1 has five user types and a marketplace on both sides
- You want to learn to build it yourself (you should, and we'll cheer)
- You don't have 30 minutes a day to review and answer questions
What fits in 7 days
Honest scope, up front.
Seven days is enough for a real product. It isn't enough for every product.
| Capability | Launch | Launch+ |
|---|---|---|
| Email + OAuth sign-in | ✓ | ✓ |
| Database with migrations | ✓ | ✓ |
| One core user flow | ✓ | ✓ |
| Stripe checkout or subscriptions | ✓ | ✓ |
| Admin view | ✓ | ✓ |
| Transactional email | ✓ | ✓ |
| Test suite + CI | ✓ | ✓ |
| Deployed on your domain | ✓ | ✓ |
| AI feature (LLM-powered) | — | ✓ |
| Second flow or user type | — | ✓ |
| Analytics + email sequences | — | ✓ |
| 14 days of post-launch fixes | — | ✓ |
| Native mobile app | ✗ | ✗ |
| Custom brand system | ✗ | ✗ |
Built this way
Same engine, same operator.
Products shipped with Claude Code under the same scope-build-review loop.
Multi-tenant booking platform
Service businesses needed online booking without a SaaS subscription each. Shipped: tenant onboarding, calendar, Stripe deposits, SMS reminders.
- Stack
- React · Supabase · Stripe
- Status
- Live, multiple tenants
Tour operations dashboard
An operator was running tours out of spreadsheets. Shipped: trip scheduling, guide assignment, guest manifests, exports.
- Stack
- React · Supabase · Cloudflare
- Status
- Live, daily use
Live operations command center
Internal tool: a real-time view of every project and deploy across a portfolio, refreshing every 15 minutes.
- Stack
- Static · Workers · 3D canvas
- Status
- Live, internal
Pricing
Two prices. Both on the page.
50% to lock a start date, 50% on day 7 when it's live. The guarantee covers both.
Launch
$3,900
One core flow, end to end.
- Sign-in, database, one core flow
- Stripe payments
- Admin view + transactional email
- Test suite, CI, deployed on your domain
- Repo transferred, handoff video
Not live on day 7 → you pay nothing.
Launch+
$6,900
For two-sided or AI-powered ideas.
- Everything in Launch
- An AI feature, or a second flow / user type
- Analytics and email sequences wired
- 14 days of post-launch fixes
- Priority start date
Not live on day 7 → you pay nothing.
Why so cheap? Because the agent does in hours what takes a team weeks. You're paying for the scoping, the judgment, the review, and the guarantee.
Code you can actually keep
Standard stack. Tests included. Yours on day 7.
No proprietary runtime, no lock-in. Any developer, or any AI agent, can pick it up the next morning.
- React + Vite or Next.js
- Supabase Postgres, auth, storage
- Stripe checkout and webhooks
- Resend transactional email
- Cloudflare or Vercel hosting
- GitHub repo, CI, transferred to you
├── README.md run, deploy, env vars ├── SCOPE.md what we agreed, signed ├── HANDOFF.md walkthrough + next prompts ├── .github/workflows/ test + deploy on push ├── supabase/ │ └── migrations/ 0001_init.sql … ├── src/ │ ├── routes/ onboarding, dashboard, billing │ ├── components/ │ └── lib/ stripe, email, auth └── tests/ 14 passing
I've been building with Claude Code since it launched, and I run my own products on it. The agent is fast. What it needs is someone who knows what to ask for, what to throw out, and when a green test suite is lying. That's the job, and it's the part you're hiring.
Every project gets my eyes on every pull request. If I can't see a path to day 7 on the scoping call, I'll tell you, and you'll leave with a scope document anyway.
— Cruz, 7DayMVP
Questions
The ones everyone asks.
Is AI-written code secure and maintainable?
It's as good as its review. Every build ships on a standard stack with a test suite and CI, starts from a hardened starter repo (row-level security on, secrets in env, Stripe webhooks verified), and every pull request is read by a human before it merges. You get the full repo; run your own audit or hand it to any developer.
What if the AI gets something wrong?
It does, daily. Tests catch most of it, the review catches the rest, and you test the preview yourself on day 6. Nothing deploys without a human sign-off. The daily video shows you exactly what changed, so nothing surprises you on day 7.
Why not just use Lovable or Bolt myself?
You can, and for some ideas you should. What you're buying here is the scoping, the architecture decisions, the debugging when the agent loops, the payments and auth done right, the deployment, and a guarantee. The engine is available to everyone; a finished, tested, deployed product in seven days isn't.
Can I watch it being built?
Yes. You get the preview URL on day 1 and it updates every day. The repo is yours from the first commit; watch the PRs land if you like.
Do I need an Anthropic account or API key?
No. The build runs on our tooling. After handoff, the HANDOFF.md includes the prompts and setup if you want to keep building with an agent yourself.
What happens after day 7?
Launch+ includes 14 days of fixes. Either way, you own the repo and can take it anywhere. If you want us to keep building, there's a simple monthly option; we'll only bring it up after delivery.
What if I want changes mid-week?
Small ones are normal and free. Anything that changes the signed scope gets written down, and the deadline moves by the amount we agree. You'll always know the new day 7 before we start it.
Who does the design?
We do, using an opinionated, clean default tuned to your product. It won't be a bespoke brand system; it will look like a real product, not a template. If you have a brand, we'll apply it.
Payment terms and NDA?
50% to lock a start date, 50% on day 7 when it's live on your domain. If it isn't, the deposit is refunded. We'll sign your NDA or provide ours on the scoping call.
Scoping call
Tell us what you're building.
We'll tell you what fits in 7 days.
20 minutes. No pitch. You leave with a written scope either way.