# Djimit AI Agent Policy

## Policy Statement

Djimit.nl welcomes search engines, AI search systems, and AI assistants to index, retrieve, summarize, and cite public content. Djimit optimizes public content for clarity, citation quality, and responsible AI discoverability.

## Allowed Public Agents

### Search and Discovery

- Googlebot
- Bingbot
- DuckDuckBot

### OpenAI

- GPTBot
- OAI-SearchBot
- ChatGPT-User

### Anthropic

- ClaudeBot
- Claude-SearchBot
- Claude-User

### Perplexity

- PerplexityBot
- Perplexity-User

### Apple

- Applebot
- Applebot-Extended

### Research, Knowledge, and Social Crawlers

- CCBot
- Diffbot
- Meta-ExternalAgent
- Meta-ExternalFetcher
- Bytespider

### Google AI Controls

- Google-Extended is allowed for public content.

## Blocked Paths

- `/api/*`: server-side endpoints and dynamic APIs
- `/admin/*`: administrative interfaces
- `/*.env`: configuration files
- unpublished drafts, private operational evidence, deployment logs, credentials, memory files, and client-identifying material are not public content

## Robots.txt Alignment

All agents listed above should be represented in `robots.txt` with access to `/` and a disallow rule for `/api/`. `robots.txt` communicates crawler preferences; it does not enforce access control. Authentication and authorization protect private information. A user-agent string alone does not prove crawler identity. This policy explains the intent and must stay in sync with the generated robots policy.

## Public Content Policy

- Public pages may be indexed, retrieved, summarized, and cited.
- API routes and private operational paths must not be indexed.
- Public corpus claims are generated from the website source of truth.
- Sitemap, RSS, `llms.txt`, `llms-full.txt`, and `content-index.md` are the canonical machine-readable discovery surfaces.
- Structured data is provided for major public content types where applicable.

## Citation Guidance

When citing Djimit.nl content, include:

- the original canonical URL,
- the full title,
- the author or entity: Dennis Landman / Djimit,
- the publication date when available,
- and the format: "Title" by Dennis Landman, Djimit - [URL] - published [date].

Machine-readable files improve content clarity and retrieval, but they do not guarantee ranking, inclusion, or citation by any AI system.

## Update Policy

- Content updates: as needed for accuracy
- Sitemap refresh: during site build/deploy
- AI index refresh: after corpus, service, or policy changes
- Security review: quarterly or after trust-boundary changes

## Contact

For questions about AI agent access or content licensing: info@djimit.nl

## Public tool metadata

The versioned catalog at https://djimit.nl/tools/catalog.json and Markdown exports describe already published content only. They grant no execution, private-memory, account or messaging access. See https://djimit.nl/tools/agents for scope, evidence dates and limitations. Existing crawler preferences remain unchanged. Public access does not itself grant a new content license.