Use the docs with AI
Hand a page or the whole docs to an assistant or a coding agent.
These docs are written to be read by people and by assistants alike. Hand an assistant or a coding agent what it needs, and ask it to build your integration from there.
One page
Every page has the same row under its title:
| Action | Does |
|---|---|
| Copy page | Puts the page on your clipboard as Markdown, to paste into any assistant |
| Markdown | Opens the page as Markdown: its address is the page's with .md at the end |
| Open in ChatGPT | Starts a chat that asks ChatGPT to read the page's Markdown |
| Open in Claude | Starts a chat that asks Claude to read the page's Markdown |
The whole docs
| File | Holds |
|---|---|
/llms.txt | An index of every page's Markdown, by section, as llmstxt.org proposes |
/llms-full.txt | Every page in one Markdown file |
Give a coding agent /llms-full.txt for the whole picture.
A prompt to start from
Put this site's address in place of <this site's address>, and your stack in place of the other:
Read <this site's address>/llms-full.txt, the docs of the YOU(th) webhook and read API.Build a webhook receiver for my backend in <your language and framework> that:- verifies each delivery's X-youth-signature with the secret in YOUTH_WEBHOOK_SECRET, over the raw body, and refuses one older than five minutes;- skips an event_id it has already handled, and answers 200 within ten seconds;- for Checkup.completed, reads the bundle at resource_url with an OAuth2 client-credentials token from YOUTH_CLIENT_ID and YOUTH_CLIENT_SECRET, reusing the token until it expires.Write tests for each of these, and keep every secret out of the code.What these pages describe is a preview: tell your assistant so, and check its work against them.