AI / LLM integration

This documentation is built LLM-first: every page is real server-rendered HTML backed by a markdown source, and that source is exposed directly so language models and agents can consume it without scraping rendered HTML.

The surfaces

Surface	What it is
/llms.txt	An llmstxt.org-format index — the site summary plus a link to every page's raw markdown, grouped by section.
/llms-full.txt	The full corpus: every page's markdown concatenated into one file.
/docs/<slug>.md	The raw markdown for any page. Append .md to a docs URL, or use the "Copy page" control's View .md.
Copy page control	The button at the top of each page copies that page's exact markdown source to your clipboard.

All of these are generated from the same markdown files that render the pages, so they never drift from what a human reads.

Pointing an agent at the docs

• For a lightweight index an agent can crawl, start from /llms.txt and fetch the per-page .md links it needs.
• For a single-shot context load, fetch /llms-full.txt — it contains the whole documentation set.
• For one page, fetch its .md URL (for example, https://app.gma.fi/docs/products/gmstbl.md).

The raw markdown uses standard headings, tables, and fenced code blocks, so it maps cleanly into a model's context. Diagrams are referenced as images with descriptive alt text, which is preserved in the markdown.

Building on the API instead?

If you want live data rather than documentation, see the API reference — it is Bearer-key gated and returns JSON.

Next: Community & support