Which search index feeds ChatGPT, Gemini, Copilot, Claude?

There is no single answer: each AI product gets its search results from a different place, and most companies do not say so plainly. The table below separates what is official, what comes from secondary sources, and what we could not verify, based on documents we read on 20 October 2026.

Which search index does each AI product use?

The table is a snapshot; products change. Details for each row follow.

What do Google's AI Overviews and Gemini use?

Both use Google's own index. Google's AI features documentation (last updated 10 December 2025) says that to be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet, and that there are no other technical requirements. So there is no separate registration or file for AI summaries; being in the index is the condition, and being in the index is not guaranteed.

For Gemini, Google's crawler documentation (last updated 15 October 2026) explains the Google-Extended control by tying "grounding", meaning providing content from the Google Search index to the model at prompt time, to Gemini Apps and to Grounding with Google Search on Vertex AI. The same page says Google-Extended does not affect a site's inclusion in Google Search and is not a ranking signal.

Does ChatGPT take its search results from Bing?

We found no sentence in OpenAI's documentation that says so. What we know:

• OpenAI's crawler page says OAI-SearchBot is used to surface websites in search results in ChatGPT's search features. Sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though they can still appear as navigational links. A robots.txt change can take about 24 hours to reach their systems.
• The developer documentation for ChatGPT's web search tool does not name an index provider either.
• OpenAI's ChatGPT search help page and its announcement post were not reachable for us (the server returned 403), so we could not verify what they say.
• Secondary sources: InfoQ (5 November 2024) writes that internet searches in ChatGPT search are performed by third-party engines, including Bing, and that OpenAI partnered directly with some publishers. Yoast (12 November 2024, updated 10 December 2024) says it uses Bing's index and other sources.

These are almost two years old and secondary. We could not confirm with OpenAI that they still hold. The safe conclusion: allow OAI-SearchBot and make sure your pages are also in Bing's index, but there is no official document saying that being in Bing gets you into ChatGPT.

What does Perplexity use?

Perplexity's bot page says PerplexityBot "is designed to surface and link websites in search results on Perplexity" and is not used to collect content for model training. Perplexity-User is a separate identity that can visit a page when a user asks a question; the page says it generally ignores robots.txt rules because the user requested the fetch. Whether Perplexity also uses another search provider's results is not stated in the documents we read: not verified. We could not reach Perplexity's blog post (perplexity.ai/hub/blog/introducing-the-perplexity-search-api), its FAQ page (perplexity.ai/hub/faq/how-does-perplexity-work) or the research.perplexity.ai article from our environment (403), so we did not verify what they say. A reviewer could open the perplexity.ai/hub/blog index separately, but we did not look in it for this topic.

What is the link between Claude and Brave?

Anthropic's support page lists three bots: ClaudeBot (collects content that may be used for training), Claude-User (visits sites when a user asks Claude a question) and Claude-SearchBot (navigates the web to improve search result quality). It says that disabling Claude-SearchBot prevents Anthropic's system from indexing your content for search, which may reduce your visibility in search results.

For Brave there is an official trace. Anthropic's subprocessor list (its Trust Center) shows "Brave Search" for the function "Web Search", for all products, based in the United States. We opened the list in a browser on 20 October 2026; the page text does not arrive in the first HTML and is loaded by JavaScript. The list shows that Brave processes data for web search. Anthropic does not say that all of Claude's search results come from Brave's index, or what exactly Claude-SearchBot does in this.

Secondary sources are older but point the same way: Brave Search was added to the list on 19 March 2025 (Simon Willison, 21 March 2025), a parameter called "BraveSearchParams" was seen in Claude's search function, and one search's citations matched Brave's results exactly. TechCrunch (21 March 2025) wrote that Anthropic "appears" to use Brave.

What does this mean for your site?

There is no single index, so there is no single check. Start with what is documented:

1. Be in Google's index. It is the shared condition for AI Overviews, AI Mode and Gemini. Watch the status in Google Search Console.
2. Be in Bing's index too. It is the documented route for Microsoft 365 Copilot, and secondary sources name Bing for ChatGPT. Watch the status in Bing Webmaster Tools.
3. Allow the search bots in robots.txt. The ones named in this article: Googlebot, Bingbot, OAI-SearchBot, PerplexityBot, Claude-SearchBot. Our article on which to allow and why: Robots.txt for AI bots.
4. Check that your CDN or firewall does not cut the bot. A bot you allowed in robots.txt can still be stopped at the door: Is Cloudflare blocking AI bots you allowed in robots.txt?
5. Verify real bots in your log. A name in the log can be written by anyone: AI bots in your server log.

Ruescan is Rueseo's site scanner; one of its 275 signals looks at the permission robots.txt gives AI bots. Of the 30 Turkish sites we measured on 19 October 2026, 4 blocked at least one of the search bots that collect sources for AI answers in robots.txt (details and limits of the measurement).

Frequently asked questions