Botservatory

Catalogue

Who is out there

34 identities from 15 operators, checked against what each operator publishes. Where no operator documentation exists, the entry says so.
Training
Collects pages for model training. Sends nobody back.
Search index
Builds an index that answers cite. The class that can send visitors.
User fetch
Fetches one page because a person asked. Usually skips robots.txt by design.
Agent
Browses and acts on a person's behalf.
Control token
A robots.txt token only. No crawler ever sends this name.
Link preview
Fetches a page to render a link card.

OpenAI

Anthropic

Google

Microsoft

Perplexity

Meta

Amazon

Apple

Common Crawl

DuckDuckGo

Mistral

ByteDance

Cohere

Diffbot

xAI