# For agents

How an LLM or a script should read these docs and build on Ripping.

These docs are written to be read by software as well as people.

## Machine-readable copies

| Address | What it is |
|---|---|
| `/llms.txt` | A short index: every page with its title, one-line description and address. Start here. |
| `/llms-full.txt` | Every page, in full, in one file. About the size of a long README. |
| `/<page>.md` | Any page as raw markdown, e.g. `/build/payments.md`. Same content as the HTML page. |

All three are served from `https://docs.getripping.com` and need no sign-in.

## What a generated app is

A Ripping app is a Vite + React project. Its own code lives in `src/`; the modules under `src/lib/` are the whole surface an app has for accounts, data, files and payments, and they call Ripping's API on the app's behalf. The **Building apps** section documents each module with its real signatures. When you write code for a Ripping app, import from those modules rather than calling the HTTP routes directly; the routes are documented so you know what the modules do, not as a second way in.

## What not to do

- Do not send amounts, prices or keys from the browser. Prices come from the owner's catalogue; keys are held by Ripping.
- Do not build a sign-in, a database layer or a payments flow from scratch. Each exists already and is the only one the platform will talk to.
- Do not remove the Settings or Purchases screens from an app's route table. Restyle them if you must.
