Google describes this token as controlling specified Gemini training and grounding uses. Grounding means using retrieved content to help support an answer. The token does not control ordinary Google Search inclusion or ranking. (Google common crawlers documentation ↗)
Jump to a section
Keeping Search separate from Gemini use
A publisher considers the Google-Extended token for specified Gemini-related uses and checks the corresponding line in robots.txt. It separately checks the page’s Google Search access because Google says this token does not change Search inclusion or ranking.
A token, not a separate crawler
There is no separate Google-Extended request user-agent to identify in a server log. It is a product-use control applied through robots.txt. Inspect Googlebot access separately when diagnosing why a page is missing from Google Search.
Test the policy without conflating it with indexing
Save the published rule and a representative public article. Record the intended Google product use separately from ordinary Search access. If search visibility later changes, also check redirects and crawler access before attributing the change to the policy edit. Keep the previous rules in the change record so you can identify exactly which instructions moved.