Free tool
Is your AI-built app production ready?
You proved the idea with a prototype built in Bolt, Lovable, Cursor, or Claude. Before real users hit it, score it. Twelve questions across the six things that separate a demo from software that survives Customer Two. You'll get an instant score and see the one axis to fix first. No email required to see your result.
Six axes. Fifteen quick questions. About two minutes. Tap to answer; honest beats generous, and an "I don't know" counts against you on purpose.
1 of
Before you see it, what do you think you scored out of 100?
No wrong answer. It just makes the number mean more when you see it.
Not your score
How we worked this out
Your score is the weighted average of the six axes. Each axis is scored from your answers, then multiplied by its weight. The composite is that weighted total, rounded to the nearest whole.
| Axis | Score | Weight | Points |
|---|---|---|---|
| Composite |
Fix this first
Move one lever
Your saved link and card always keep your real answers.
Send this to whoever signs off
Or save the card to paste into a deck or an email
Nothing you enter leaves your browser unless you ask us to email your report. The score is computed on your device.
The six axes, and what breaks without them
A prototype is built to prove an idea quickly. Production software is built to survive real users, real data, and real failure. These are the six places where that gap shows up. The score weighs them the way production does: your data model and your permissions carry the most.
- Data model (25%)
- Weak here: Records collide, duplicate, or quietly overwrite each other.
- Auth and permissions (20%)
- Weak here: One user can open another user's data by changing a URL.
- Concurrency (15%)
- Weak here: Two people edit one record and one of them loses their work.
- Errors and observability (15%)
- Weak here: It breaks at 2 am and a customer tells you before your logs do.
- Migration path (13%)
- Weak here: You cannot change the schema without downtime or losing data.
- Audit trail (12%)
- Weak here: You cannot prove who changed what, or when, after the fact.
We rebuild AI-built and no-code prototypes into database-first production software. If your score says the foundations aren't there yet, that's the work we do. See how prototype-to-production works.