TheyfetchUnderstand your data
Useful evidence. Less data.
Request metadata helps identify AI fetchers. The analytics keep the result of that classification, without raw IPs or full user-agent strings.
What happens to a page request
- IP address
- Used to verify the fetcher against supported operator rules, then discarded before analytics storage.
- Full user agent
- Used to match the bot catalog. The matched bot token and purpose are retained instead of the raw string.
- AI page requests
- Recognized AI requests are retained with their path, bot, purpose, and verification status. Human browsers and unknown crawlers are dropped.
- Request bodies
- Request bodies are never logged on the ingest path.
What prompt tracking keeps
Your workspace retains the questions you configure, answer text, citations, search queries, named brands, model, surface, and cost. These records let you inspect the evidence behind a result.
Services that support Theyfetch
Cloudflare runs the app and transactional storage. Tinybird stores citation analytics in the EU. Resend sends email, and Stripe handles billing. Provider APIs process the prompts you enable. OpenRouter routes AI requests to the selected model provider, including answer text for extraction and product information for setup suggestions.