The catalog
Six native formats.
One substrate.
Each profile is a published contract: a name, a media type, a sample, and a coherence window. The signed catalog is at /.well-known/crawler-formats.json.
OpenAI · GPTBot
JSON-LD (schema.org Dataset)
Anthropic · ClaudeBot
Markdown with semantic anchors
Google-Extended
structured-data-first HTML
Perplexity · PerplexityBot
Markdown with explicit sources
Common Crawl · CCBot
Archival HTML with metadata
Token-holder · negotiated
JSON-LD · resource map · RDF · raw 13D
Catalog
The catalog is signed.
The catalog endpoint exposes the same six profiles in machine-readable form, with a detached signature and a manifold-state hash.
A consumer that wishes to verify the catalog independently may fetch /.well-known/crawler-formats.json and check the embedded signature against the published trust key. The catalog is re-signed on each coherence cycle.