For AI coding agents

These docs publish an llms.txt at the site root: a single plain-text index of every page and what each one covers, plus the facts an agent most often gets wrong about Tangent.

https://<docs-host>/llms.txt

Point your agent at that file rather than at the rendered HTML. It is short enough to fit in a context window whole, and it is generated from the same content as the site.

What it contains

• The data model in a paragraph: events, deviceId versus userId, properties versus user properties, groups, entities, revenue.
• The delivery guarantees: per-event validation, at-least-once with messageId de-duplication, and quarantine rather than silent drops.
• One page per line, with a description specific enough to route a question to the right page.
• The publication caveat. No SDK is on a public registry yet, so an agent that generates npm install @tangent/analytics-node or pip install selica-analytics has produced a command that fails. llms.txt states this near the top for exactly that reason.

Keeping it accurate

llms.txt lives at docs-site/content/public/llms.txt and is copied to the site root by the content sync step. When you add a page to the sidebar, add its line here in the same change — nothing generates it for you.