# What Bots Does answer.cloud Track?

> Canonical source: [What Bots Does answer.cloud Track?](https://www.answer.cloud/pages/what-bots-does-answer-cloud-track-hosted/)

Understand agent analytics categories, how requests are observed, and why browser analytics cannot identify every crawler.

answer.cloud agent analytics categorizes observed requests, including ordinary visitors, search crawlers, user-triggered AI fetchers, and model-training bots where the request can be classified.

## Where the evidence comes from

Request-level analytics at the server or edge can observe a request even when it does not run JavaScript. Browser analytics records browser-side activity and is useful for human sessions and referral behavior, but it is not a complete crawler log.

For an existing Cloudflare-proxied website, use the [Worker integration](https://www.answer.cloud/docs/cloudflare-worker-agent-analytics-hosted/). A Worker must be attached to the relevant website routes; a standalone workers.dev page cannot observe requests to your domain.

## Interpret classifications carefully

Provider user agents and their purposes can change. A user-agent string can also be spoofed. Treat dashboard categories as classifications of observed requests rather than proof of a provider’s identity or a complete list of every AI bot on the internet. Use published provider verification information where available.

Separate training crawls, search indexing, and user-triggered fetches when the data supports that distinction. A crawler request does not establish a citation, a hidden search query, or the content of a visitor’s separate AI conversation.

Pair crawl activity with human AI referrals, on-site chat questions, and controlled answer tests. See [visibility metrics](https://www.answer.cloud/pages/llm-visibility-metrics-the-kpi-playbook-for-seo-and-marketing-leaders-hosted/).