Free and MIT
The Next.js boilerplate your agent already knows.
Sign-in, billing and admin already work. Rules and guard hooks keep your coding agent on track from the first prompt.
- Boots with no keys
- Every combo tested
- Free and MIT
No card. We email you the link.
Bash(rm -rf .next node_modules)
BLOCKED by block-destructive
recursive force delete (`rm -rf`)
Do this instead: delete the specific paths you mean, without -f, or move them aside (`mv <path> /tmp/`).
26 batteries. Pick your stack. One per category, or none.
A blank repo makes your agent guess. This one tells it.
| Blank repo | Agentic Boilerplate |
|---|---|
| Re-explain your stack every session | Agent reads CLAUDE.md first |
| A new pattern in every file | Rules load for the files it edits |
| Same Stripe bug, fixed three times | Known fixes, written down |
| rm -rf runs. Work gone. | A guard hook blocks it first |
| Days wiring auth, billing, admin | All three work on first boot |
Sign-in to admin. Already wired.
Real screenshots of a generated app. Nothing mocked up.
Every way in, done
- Email, password, magic links
- Google, GitHub, Microsoft
- Better Auth, Clerk or Supabase Auth
A signed-in app, not a stub
- Sidebar, account menu, settings
- Profile and security pages
- Your pages drop right in
Take money on day one
- Monthly, yearly and one-time
- Dodo Payments, Lemon Squeezy, Polar or Stripe
- Pricing renders with no keys
Run it without a SQL shell
- Search every user
- Ban, impersonate, audit log
- Every action is recorded
A site for your product
- Hero, features, pricing, FAQ
- Terms, privacy and llms.txt
- All from one file
Your agent will try it. Your repo says no.
Pick a command an agent might run. The hook in your zip answers.
rm -rf ./.verify-hooks-probe
BLOCKED by block-destructive: recursive force delete (`rm -rf`)
A recursive force delete removes untracked files and uncommitted work with no recovery path.
Every generated repo runs these probes in its own self-test. Real refusals, word for word.
30
Path rules
Loaded only for the files it touches
7
Guard hooks
Stop rm -rf, force pushes and .env leaks
28
Skills
Slash commands for the jobs you repeat
74
Known fixes
Read before your agent guesses
Counts from the Indie SaaS preset.
Pricing? $0. Really.
No tiers, no trial, no card. The whole thing is MIT.
- Sign-in, billing and admin
- 26 batteries to pick from
- 31 designs to start from
- Rules, hooks, skills, known fixes
- Claude Code, Codex and Cursor
- Every pair tested together
- Yours to keep, no attribution
The same system ran first on Hostshare, a live client codebase: 797 commits and 131 PRs merged in about 3 weeks.
Questions. Short answers.
Is it really free?
Yes. The builder and every repo it makes are MIT. No tiers, no trial, no attribution.
Why do you want my email?
You download right away. We also email the link so you can grab it again for 7 days. Nothing else unless you tick the box.
Do I need Claude Code?
No. Claude Code gets everything, hooks included. Codex and Cursor get the same rules and skills in their own formats.
What if I pick a combo nobody tested?
You can't. Untested pairs are greyed out with the reason. Every pair you can pick was installed, typechecked, built and booted together.
Can I change batteries later?
Yes. The zip carries agentic.config.json. Regenerate with a new pick and diff the result.