# Docs for AI agents

Canonical page: https://docs.ovok.com/working-with-ai-agents/docs-for-agents

Ovok Docs publishes Markdown for AI tools alongside each rendered page:

- Start with the [docs router for AI agents](/llms.txt). It points to the most useful topic pages without loading the whole documentation set.
- The [exhaustive page manifest](/llms/manifest.json) lists every published page with its title, canonical URL, page-specific Markdown URL, section, source hash, and `lastUpdated` timestamp in UTC. It also lists topic downloads with their versions, page counts, and UTF-8 byte sizes.
- Download a [topic bundle](/llms.txt) for focused context, such as [Authentication](/llms/topics/authentication.txt), [Access policies](/llms/topics/access-policies.txt), or [FHIR API references](/llms/topics/api.txt). Each topic has its own content-based version, so unrelated documentation changes do not invalidate that download.
- For a focused task, fetch the page-specific Markdown URL, such as [Authentication](/authentication/llms.txt), [Access policies](/access-policies/llms.txt), or [Settings and features](/settings-and-features/llms.txt). The same pattern works for every docs page: `/path/to/page/llms.txt`.
- Every page is also served as Markdown at `/path/to/page.md`, or at its own URL when the request sends `Accept: text/markdown`.
- Use the [complete docs bundle](/llms-full.txt) only when broad context is needed. It contains every page and is intentionally the largest option; for most implementation tasks, start with the router and fetch only relevant pages.

Bundle versions are SHA-256 identifiers of the exported text before its version line is inserted. Compare them to detect changed downloads. File sizes are not token counts; if a topic still exceeds your model's context budget, fetch only the page-specific Markdown you need. Page timestamps record source-content changes rather than rebuild or deployment times.

The [Ovok MCP server](/working-with-ai-agents/mcp-server) points connected agents to the router too.

For example, you can ask: “Read the Ovok docs index, then fetch the page-specific Markdown for authentication, FHIR resources, and settings that apply to this workflow. Use those docs as the source of truth and cite the relevant pages in your plan.”