llms.txt

This documentation is published in a form assistants can read directly, without any integration.

The files

/llms.txt - an index. Every page, with its title, a one-line description and its URL. Small enough to drop into a prompt whole.

/llms-full.txt - the entire documentation as one plain-text file. Larger, but complete: an assistant that reads it has the whole product.

Both follow the llms.txt convention.

Using them

Point any assistant at the URL:

Read https://docs.orijin.io/llms-full.txt and then tell me how to run a deforestation check on a plot.

Or paste llms.txt into a system prompt, and let the assistant fetch the specific pages it needs.

Why this exists

These pages are the single source of truth for how Orijin works. The same Markdown produces this website, the help library behind the in-dashboard AI Assistant, and these files.

That’s deliberate. Documentation that only humans can read goes stale, because nobody notices. Documentation that an assistant answers from is exercised every day, by every user who asks it a question - and a wrong answer gets reported.

It also means an assistant working with Orijin doesn’t have to guess at screen names or invent buttons. It reads what’s here.

For engineers changing Orijin

If you’re an agent or a developer changing a user-facing flow, the doc page is the thing you update - not the generated help library, and not this file. See docs-site/CONTRIBUTING.md in the repository.