Gemini-Deep-Research
Fetches sources when a Gemini user runs Deep Research. Tracked by crawler registries but absent from Google's crawler docs.
Reference reviewed by Sona
The operator publishes little or no documentation for this token. Details here come from community crawler registries and observed behavior, not vendor confirmation.
| Operator | |
|---|---|
| Powers | Gemini's Deep Research reports |
| Purpose | User-triggered fetch |
| User-agent token | Gemini-Deep-Research |
| Respects robots.txt | No (user-triggered) |
Gemini-Deep-Research is associated with Deep Research, the Gemini feature that commissions a long-form report and gathers the sources behind it. Where an ordinary fetcher retrieves one page, a Deep Research run works through many - so the characteristic signature is a burst of multi-page fetches traceable to a single user request.
A provenance caveat matters here. This token is tracked by crawler registries including Dark Visitors and CrawlerCheck, and Deep Research itself is documented in Google's Gemini API materials - but the token does not appear in Google's own crawler documentation, neither in the user-triggered fetchers list nor among the special-case crawlers. So the identifier is community-tracked rather than officially published, and any statement about its robots.txt behavior is inference from how Google treats its documented fetchers.
That inference is reasonable but should be labelled as such. Google's documented user-triggered fetchers "generally ignore robots.txt rules" because a person initiated the request, and Deep Research plainly fits that pattern. Plan for a robots.txt Disallow to be ineffective, and use server- or firewall-level filtering if you need the block enforced.
The exposure is unusually valuable if you allow it. Deep Research output is long-form, heavily cited, and produced for someone who asked a substantial question and intends to act on the answer - among the highest-quality citation contexts available in any AI product.
How Gemini-Deep-Research behaves
- Produces bursts of multi-page fetches from a single user request, unlike single-page fetchers.
- Not listed in Google's crawler documentation - the token is community-tracked.
- robots.txt behavior is inferred from Google's documented user-triggered fetchers, not stated for this token.
- Deep Research output is long-form and heavily cited, so an allowed fetch can yield prominent attribution.
Allow Gemini-Deep-Research
Deep Research reports can read and cite your pages - high-intent exposure inside long-form research that users actually act on.
User-agent: Gemini-Deep-Research Allow: /
Block Gemini-Deep-Research
You don't want burst research traffic - but a robots.txt rule alone probably won't stop it, so enforce it in your WAF.
User-agent: Gemini-Deep-Research Disallow: /
Heads up: Gemini-Deep-Research does not reliably honor robots.txt. To enforce a block, combine the rule above with server- or firewall-level filtering of the user-agent.
How to verify Gemini-Deep-Research
Published IP range feed
Because Google does not document this token, there is no feed published specifically for it. Google's user-triggered-fetchers IP range file is the closest applicable source, and reverse DNS under google.com applies to Google infrastructure generally - but neither confirms this particular identifier. Verify that the traffic is Google's, and treat the specific token as unconfirmed rather than authenticated.
Documented reverse-DNS suffixes: .google.com
Commonly confused with Gemini-Deep-Research
Both are Gemini-driven, but Google-Agent performs actions while Deep Research gathers sources for a report. Google-Agent is officially documented; this token is not.
Notebook fetches documents a user chose to study; Deep Research discovers and fetches sources itself across many pages.
Also absent from Google's crawler documentation, and also Gemini-related - but URL-context serves API developers supplying specific URLs.
Gemini-Deep-Research FAQs
Is Gemini-Deep-Research officially documented by Google?
No. The token is tracked by third-party crawler registries, and the Deep Research feature is documented in Google's Gemini API materials, but the user-agent itself appears in neither Google's user-triggered fetchers list nor its special-case crawlers list.
Will robots.txt stop it?
Probably not, though this is inference rather than documentation. Google's documented user-triggered fetchers generally ignore robots.txt because a person initiated the request, and Deep Research fits that pattern. Use edge filtering if the block must hold.
Why am I seeing many fetches at once?
That is the expected signature. A single Deep Research request gathers sources across many pages, so a burst of fetches from one user action is normal rather than a runaway crawl.
Is being cited in Deep Research valuable?
Unusually so. The output is long-form and heavily cited, produced for someone who asked a substantial question and intends to act on it - a higher-quality citation context than a quick chat answer.
Can Gemini-Deep-Research read your page right now?
Test any URL and see exactly what AI crawlers receive.