payo

Generate project-tailored AI assistant rules & skills in minutes.

Payo interviews you about your stack, then writes the guidance files your AI coding assistant reads — so Claude, Cursor, Copilot, and friends follow your project's conventions instead of guessing.

---

Contents

• What is Payo?
• Why Payo?
• How Payo differs from a generic skills pack
• Who is this for?
• Quick Start
• How to use — a walkthrough
• Already have a project? Payo detects your stack
• What gets generated
• What it asks about
• AI vs. template generation
• Bootstrap prompt
• Resume anytime
• Roadmap
• Requirements
• Run locally
• Contributing
• License

What is Payo?

AI coding assistants are only as good as the context they're given. Payo is an interactive CLI that interviews you about your project — language, framework, database, auth, testing, conventions — and then generates the guidance files your assistant actually reads.

It supports Claude, Cursor, GitHub Copilot, Codex, Windsurf, Zed, and Antigravity. Where the assistant ships a headless CLI, Payo drives that tool's own AI to write rich, project-specific docs; where it doesn't, Payo falls back to solid templates — so you always end up with usable output.

And for those wondering what Payo means, in Gujarati it means foundation — what a building or project is built on.

Why Payo?

Every AI coding session starts cold. Your assistant doesn't know you use Drizzle over Prisma, that handlers live in src/routes, that you write Vitest specs beside the file, or that commits follow Conventional Commits. So you re-explain it — in chat after chat — and still get code that ignores half of it.

The fix is the guidance files agents read — an AGENTS.md entrypoint plus Agent Skills under .agents/skills/. But writing them by hand is tedious and easy to get wrong. Most people never do it, or do it once and let it rot.

Payo writes them for you in minutes — tailored to your actual stack, in one universal layout that every skills-compatible tool reads (Claude Code, Codex, Cursor, Copilot, Gemini, Antigravity, Windsurf, Zed, …) — so your assistant follows your conventions from the first prompt instead of guessing.

Vibe coding, without the mess

Vibe coding is fast and fun — you prompt, the AI improvises, code appears. The problem isn't the speed, it's that the assistant has no rules to improvise within. So it guesses: a different ORM than the rest of the repo, files scattered wherever, its own naming, your tests and commit conventions skipped. Fine for a throwaway demo; on a real codebase it quietly piles up inconsistency you pay for later in reviews, bugs, and refactors — and each new chat starts the guessing over.

Payo gives that improvisation a guardrail. It generates the guidance files your assistant already reads, so the AI reaches for your framework, your folder layout, your testing and git rules every time — without you stopping to explain them. You keep vibe coding at full speed; the output just fits the project instead of fighting it.

How Payo differs from a generic skills pack

There are plenty of ways to get skills and rule files into a repo: curated collections, awesome-lists, starter templates, a rules pack someone published. They all hand you the same markdown everyone else gets — written for a generic project, then left for you to adapt. Payo isn't a catalog you install. It's a generator that writes your skills, from your repo.

1. Your skills are written, not downloaded. Payo doesn't ship skill content. It runs the agent CLI you already have installed and authenticated — claude, codex, cursor-agent, copilot, agy — over your answers, several skills at a time, retrying and logging each attempt. The result is authored for this project on this machine, under your account.

2. Only the skills your project needs exist at all. Every skill is gated on your answers: no database means there is no data-layer/ skill — not a stub, not an "N/A" section, no directory. Same for auth, state management, API conventions, logging, testing. A pack gives you all of them and expects you to delete the two-thirds that don't apply; Payo never creates them. The interview works the same way: a bank of 200+ stack-tailored questions across 100+ tech modules, from which you're asked only the handful your stack pulls in. Payo also shows you the computed list before writing anything, so you can uncheck any skill you don't want a file for.

3. Real commands and real paths, not placeholders. Because Payo knows your package manager, framework, and ORM, the guidance says pnpm exec eslint . — not npx eslint ., and not <your lint command>. It says cargo test, php artisan migrate, bin/rails db:migrate, and lists your monorepo's actual package directories. The generation prompt explicitly forbids guidance for any tool that isn't in your stack, so nothing generic leaks back in.

4. It reads your repo, including habits you never wrote down. On an existing project, Payo detects your stack from the manifests and infers your branch-naming and commit-message conventions by parsing your local git history (which never enters an AI prompt). In Detect everything mode your code is the source of truth: anything Payo can't find is skipped, not filled with a plausible default — so it won't prescribe /v1 API versioning, a testing setup, or a folder layout your project never adopted.

5. Guidance that's enforced, not just suggested. A skills pack is prose, and an agent can skip pr