An AI-guided interview that turns a rough project idea into a polished CLAUDE.md or AGENTS.md scaffold -- with tasks, deployment steps, and wiki pre-filled. Drop the zip into a new repo, tell Claude Code or Codex to “review project instructions”, and get to work.
Request founder pricing. Solo $39/mo or Pro $99/mo, locked in for life. Once 100 seats are claimed, prices go up to $49 and $129.
Already have an account? Sign in.
This is a live component from the actual app on an auto-loop. Each file streams in one at a time, exactly like you'll see after signing up.
Rendering your scaffold now.
Describe what you want to build in plain English. The AI asks focused follow-up questions -- stack, audience, workflow, risks. Stuck on an answer? It'll help you figure it out.
In about 60 seconds, your CLAUDE.md (or AGENTS.md), tasks/todo.md, deployment docs, and project wiki stream in live. Every file tailored to your stack. No placeholders. No TBDs.
Download the zip. Extract into a new repo. Tell Claude Code or Codex: "review project instructions." Your AI pair programmer picks up exactly where the scoping left off.
A few examples of what AI Project Scoper handles cleanly:
Claude Code projects get a CLAUDE.md. Codex projects get an AGENTS.md. Running both? You get both, and they agree with each other. No more keeping two sets of instructions in sync by hand.
Need to plan something bigger than a weekend project? Flip Super Scaffolding on the same review screen and your scoper produces a phase-by-phase implementation guide instead of a flat checklist. Every plan includes it.
A kid-friendly browser-based multiplayer pixel art co-op brawler (Gauntlet Legends meets Golden Axe) with puzzle quests, character creation, voice chat, and persistent progression for players aged 7-15.
15 files generated. Click any card above to preview, or use the Download zip button below.
Not a flat checklist. A planning artifact your AI coding tool (or your team) can execute against without missing the operational tail.
Why this phase exists, what depends on it, what comes before and after.
Concrete options, tradeoffs, and a recommendation. No “TBD”.
What can go wrong, with mitigation or “we accept this”.
Testable bullets so you know exactly when the phase is done.
Plus the operational tail most plans skip: admin panel, billing edge cases, support, abuse handling, observability, compliance — surfaced as their own phases when the project warrants them.
Solo consultants and indie builders shipping projects week after week.
Power users and agencies running multiple projects in parallel.
Standard scaffolds are unlimited on every paid plan. See full pricing & yearly options →
You can. But will you do it consistently, every project, with the same level of polish? The interview forces the structure you'd otherwise skip: audience, workflow, conventions, risks, definition of done. The output is repeatable across 50 projects instead of one good one and 49 that drift.
A zip with four files -- CLAUDE.md or AGENTS.md (depending on your tool), tasks/todo.md with project-specific phases, docs/deployment.md with real platform-specific steps, and docs/wiki.md for living decisions. Every file is tailored to the stack and scope you described.
Yes. Pick Codex during the interview and the primary doc is AGENTS.md (Codex's convention) instead of CLAUDE.md. Tasks and docs work the same way.
Right now the first 100 customers across Solo and Pro get founder pricing locked in for life: Solo $39/month or $279/year, Pro $99/month or $699/year. Once 100 seats sell, prices go up to $49/month ($349/year) for Solo and $129/month ($929/year) for Pro -- 40% off on annual either way. All plans cancel any time.