What Is IndexNow? Getting New Pages Into Bing Faster

By Muntazir · Published 6 October 2026

On 31 August 2026 this site moved from an old domain to a new one. Fifty-nine URLs on the old host got a 301 redirect each, and then sat there — a redirect does nothing for a visitor’s search ranking until a search engine actually fetches it and follows it to the new address. Left alone, that fetch happens whenever the engine’s own crawl schedule gets around to revisiting fifty-nine fairly minor pages, which can take weeks.

Instead, the same fifty-nine URLs went straight into a single request to an endpoint called IndexNow, flagged as changed. A redirect counts as a change under the protocol’s own rules, even though no content changed — the destination did. Whether that shaved real time off Bing’s recrawl is hard to isolate after the fact, since nobody ran the migration twice, once with the ping and once without. What’s measurable is the action itself, and the fact that it exists as an option at all is the more interesting part.

What IndexNow actually is

IndexNow is a push protocol. A site generates a key, hosts it as a plain text file at its own root so the mechanism can confirm it controls the domain, and then sends a URL — or up to 10,000 of them in one batch — to a shared endpoint. IndexNow’s own site lists who’s listening on the other end: Bing, Yandex, Naver, Seznam.cz and Yep. One request, several engines notified, instead of filing separately with each.

That inverts the usual relationship between a site and a search engine. Normally the engine decides when to look. A sitemap is a hint, not a summons — it tells a crawler what exists, not when to come back for it. IndexNow is the site deciding, in real time, that now is the moment to look.

The one name missing from the list

Google isn’t on it. Not quietly deprioritized, not a future rollout — absent from IndexNow’s own list of participating engines. A site still needs the basics for Google: a sitemap that resolves, a robots.txt that doesn’t block the crawler, no stray noindex tag left over from a staging environment. None of that goes away. IndexNow adds a second, faster channel for a separate set of engines; it doesn’t replace the first one, and treating it as a substitute for a working sitemap is the mistake this protocol invites.

What a submission doesn’t promise

Bing’s own setup page says this directly: using IndexNow does not guarantee a page will be crawled or indexed. It guarantees the engine was told something changed. What happens next — whether the page gets fetched promptly, whether it’s judged worth indexing, where it ranks once it is — runs through the same evaluation every other page goes through. IndexNow moves a page to the front of a queue. It doesn’t change what the queue is checking for.

Does any of this reach ChatGPT?

This is where the evidence gets thinner, and it’s worth saying so rather than rounding it up to a clean answer. For Microsoft’s own Copilot, the connection is direct: Copilot’s web answers are built from pages Bing has already crawled and indexed, and there’s no separate “Copilot index” sitting apart from Bing’s. A page that isn’t in Bing’s index isn’t available to Copilot either, so getting into Bing faster genuinely matters there.

ChatGPT is a different story, and it’s changed over the two years people have been giving SEO advice about it. Bing was, for a long stretch, central to how ChatGPT’s search feature worked — if a page wasn’t in Bing’s index, it effectively wasn’t available to be cited. That’s no longer the full picture. OpenAI built its own crawler and has been shifting away from treating Bing as its primary index, running its own retrieval alongside whatever it still draws from Bing. Exactly how much weight Bing still carries in a given ChatGPT answer isn’t something either company publishes a number for. Getting into Bing’s index remains useful. Treating it as a guarantee of a ChatGPT citation is the part that’s no longer accurate.

A smaller, odder use for the same protocol

The redirect submission at the top of this piece wasn’t pushing new content — it was telling Bing that fifty-nine old pages had changed, so it would revisit them, hit the 301s, and process the move instead of continuing to treat the old URLs as current. IndexNow covers “added, modified or deleted,” and a migration counts as modified even when every word on the page is identical to what was there before. That’s a narrower use than most explanations of the protocol mention, and one worth knowing if a domain move or a bulk URL change is coming up — the sitemap approach to the same problem works too, just on a slower clock.

What this doesn’t fix

A fast ping doesn’t make a thin page worth indexing, and it doesn’t substitute for the parts of discoverability that still run on the old rules: a sitemap Google can actually read, a robots.txt that doesn’t quietly block a crawler, pages that resolve with the right status code instead of a silent redirect chain. A sitemap validator checks the slower-moving half of this for free. IndexNow speeds up telling a search engine something changed; it was never going to be the part that decides whether the change was worth noticing. The gap between “notified quickly” and “crawled by the right bots at all” is covered in more detail in the piece on OAI-SearchBot and GPTBot and in the audit methodology.