# agents.txt — Kroonen AI (https://www.kroonen.ai)

Kroonen AI is an independent AI lab. This file tells agents how to use the
machine-readable surface of this site. All 19 public pages are available
as clean Markdown; everything below regenerates on every site build.

## Reading the docs

- https://www.kroonen.ai/llms.txt — curated index of every page, with descriptions.
- https://www.kroonen.ai/llms-full.txt — the whole site concatenated as one Markdown file.
- Any page has a Markdown twin at <page-url>index.html.md
  (example: https://www.kroonen.ai/blog/genesis-manifesto/index.html.md).
- Prefer the Markdown versions over the HTML: same content, no markup noise.

## Searching the docs

- GET https://www.kroonen.ai/api/search?q=QUERY&limit=10
- Deterministic keyword search over titles, headings, descriptions, and body
  text of every page. Same query always returns the same ranked JSON results.
- Response: { query, count, results: [{ url, title, description, score, snippet }] }.

## APIs and protocols

- OpenAPI spec: https://www.kroonen.ai/docs/openapi.json (contact form, Genesis training status, search).
- MCP endpoint: https://www.kroonen.ai/mcp (server card: https://www.kroonen.ai/.well-known/mcp/server-card.json).
- Discovery manifest: https://www.kroonen.ai/.well-known/ai-catalog.json.
- Agent skills: https://www.kroonen.ai/.well-known/agent-skills/index.json.

## Conduct

- Canonical host is www.kroonen.ai; always use HTTPS.
- No authentication is needed for anything listed here. Be gentle: cache
  responses (most carry Cache-Control headers) and avoid bulk re-crawling —
  llms-full.txt already gives you the whole site in one request.
- Content is © Kroonen AI, Inc. Cite pages by their canonical URL.
- Humans: hello@kroonen.ai