How it works

The AI reads. The code decides.

This separation is deliberate — it's what keeps recommendations consistent, defensible, and free of the 'one model recommends everything' problem.

1
Describe your idea
Type what you want to build in one plain sentence. No forms, no jargon — 'a mobile app that transcribes voice notes' is enough.
2
AI reads it into tags
A Claude model does one narrow job: turn your text into structured tags from a fixed vocabulary. It never recommends tools — it only classifies.
3
You confirm
We show what we understood as editable chips and ask the two things text rarely reveals: your country, and (if mobile) your framework. One tap to fix anything.
4
Code matches the database
Pure, deterministic code matches your confirmed tags against a hand-verified tool database — applying the regional payment rules — and ranks free-first. Same input, same result, every time.
5
Get an actionable stack
A clean, grouped stack with what's free, where free ends, the gotchas, and a copy-paste prompt to scaffold it in Claude Code or Cursor.

That's the whole thing. No account, no cost.

devAge ships production apps on exactly these stacks — region-aware payments, AI features, the works. Tell us your idea and get a plan and a quote.