How does Lalela tell a real AI agent from a fake one?
User agents are easy to fake, so every visit gets a confidence tier:
- Verified: the request came from an IP range the operator publishes (OpenAI, Perplexity, Google and Bing publish these), from a network registered to the operator, or carries a valid Web Bot Auth signature.
- Declared: the user agent names a known agent but nothing confirms it.
- Suspected: the request doesn't name itself, but its behaviour points to automation.