# Identifying AI Fetch Requests: A Grok Test

> Canonical source: [Identifying AI Fetch Requests: A Grok Test](https://www.answer.cloud/pages/grok-ai-bots-hosted/)

An observed browser-like user agent cannot establish Grok identity or the number of distinct AI bots.

A past Grok test appeared to coincide with a burst of requests to the site. That is an observation from one session, not a verified description of every Grok retrieval workflow. The older screenshot is omitted because it contains former-host addresses and does not establish reliable attribution.

## Headers are evidence with limits

Some provider agents identify themselves in a user-agent string. A browser-like string, such as a Chrome identifier, can also appear in ordinary browsers or automation. It does not prove who made the request. User-agent strings can be spoofed.

Keep request count, unique clients, and distinct verified bots separate. Several requests may come from one client. Record the test date, prompt, product settings, timing, and verification method before attributing a burst to an AI service.

For comparison, consult a provider’s published crawler information, such as [OpenAI’s bot documentation](https://platform.openai.com/docs/bots), rather than inferring identity from a browser version.

Use [request-level analytics](https://www.answer.cloud/docs/cloudflare-worker-agent-analytics-hosted/) and [defined visibility metrics](https://www.answer.cloud/pages/llm-visibility-metrics-the-kpi-playbook-for-seo-and-marketing-leaders-hosted/) together.