Is a browsing fetch the same as a crawl?

No. Crawlers visit on their own schedule to build indexes or training sets; browsing fetchers fire only when a user's request requires your page, at that moment. In logs, fetchers show irregular, human-driven timing rather than systematic sweeps.

Which user agents represent browsing modes?

The main ones are ChatGPT-User (OpenAI), Perplexity-User (Perplexity), Claude-User (Anthropic), and MistralAI-User (Mistral's Le Chat). Each vendor documents its agent strings and most publish IP ranges for verification.

What Is Web Browsing Mode in AI Assistants?

Last updated: July 5, 2026 · Menra Team

Web browsing mode is an AI assistant's ability to fetch and read live web pages on demand while answering — the user asks something, and the assistant opens specific URLs at that moment through a dedicated fetch agent. It is the mechanism behind "let me check the latest pricing" behavior in ChatGPT, Claude, Perplexity, and Le Chat.

How browsing fetches differ from crawling

The distinction is trigger and timing. Index crawlers like OAI-SearchBot sweep the web proactively; browsing agents like ChatGPT-User, Claude-User, Perplexity-User, and MistralAI-User fetch a page only because a live conversation needs it. That makes each hit in your server logs a direct trace of user demand: someone, right then, asked an assistant a question your page could answer. Several vendors state these user-triggered fetchers are not used for training, and some — Perplexity among them — argue they need not obey robots.txt because they act as a user's agent, a stance that remains contested.

Why blocking browsing agents is self-inflicted invisibility

When an assistant cannot fetch your page, it does not stall — it answers from a competitor's page or from stale parametric memory. The failure is silent: no error surfaces to you, the answer simply excludes or misdescribes your brand. Common accidental causes include CDN bot-fight modes, WAF rules that challenge unknown agents, and blanket AI-bot blocklists applied without separating training crawlers from user-triggered fetchers. Auditing this is step one in any AI mention tracking program.

What to check this week

Grep recent logs for ChatGPT-User, Perplexity-User, Claude-User, and MistralAI-User — presence means assistants are already reading you.
Confirm your WAF and bot-management rules serve these agents 200 responses, not 403s or JavaScript challenges.
Verify authenticity of heavy traffic against published IP ranges, since these strings are widely spoofed.

Example

A SaaS company sees a spike of ChatGPT-User hits on its pricing page every weekday morning — buyers are asking ChatGPT what the product costs. Because the page is fetchable and states prices in plain HTML, the assistant quotes them accurately. That is browsing mode working in your favor; the glossary entries on each individual agent cover the specifics.

Frequently asked questions

Is a browsing fetch the same as a crawl?: No. Crawlers visit on their own schedule to build indexes or training sets; browsing fetchers fire only when a user's request requires your page, at that moment. In logs, fetchers show irregular, human-driven timing rather than systematic sweeps.
Which user agents represent browsing modes?: The main ones are ChatGPT-User (OpenAI), Perplexity-User (Perplexity), Claude-User (Anthropic), and MistralAI-User (Mistral's Le Chat). Each vendor documents its agent strings and most publish IP ranges for verification.

Keep exploring

See how AI engines talk about your brand — track mentions across ChatGPT, Perplexity, Claude, Gemini and 5 more. Start with Menra