IndexNow
Last updated
IndexNow is a protocol for telling search engines about a change the instant it happens, rather than waiting for them to crawl back round and notice. When you publish, update, or delete a page, your site pings a participating engine with the URL, and the engine can fetch it promptly. It inverts the usual pull model, where discovery depends on a crawler’s schedule, into a push model where the site initiates.
It sits alongside XML sitemaps in the discovery toolkit but works differently. A sitemap is a standing list crawlers consult on their own timetable; IndexNow is an active, per-change notification. Bing’s guidance is to run both, and the two are complementary rather than alternatives.
How does IndexNow work?
The mechanics are deliberately simple:
- Generate a key. Create a unique key and host it as a UTF-8 text file at your site’s root (
https://example.com/{key}.txt) containing only the key value, so the engine can verify you control the domain. Submissions return a 403 if the file is missing or does not match. - Submit URLs on change. When a page is added, updated, or removed, send its URL to an IndexNow endpoint. A single URL goes as a GET request to
https://api.indexnow.org/indexnow?url={url}&key={key}; batches go as a JSON POST to the same host, up to 10,000 URLs per request.1 - The engine shares it. Participating engines pass submitted URLs to each other, so a single submission reaches all of them. One ping, multiple indexes.
Most managed platforms and SEO plugins can do this automatically on publish, so for many sites implementation is switching on a setting rather than writing code.
One expectation to set before you measure anything: submitting a URL is a request to crawl, not a guarantee of indexing. IndexNow shortens the gap between publishing and discovery. Whether the page then earns a place in the index still depends on the usual factors. Bing Webmaster Tools includes an IndexNow Insights report showing which submitted URLs were actually indexed and how quickly, which is the way to check rather than assuming.
Which search engines support IndexNow?
Support it: the protocol’s own participant registry lists seven endpoints: Bing, Yandex, Naver, Seznam.cz, Yep, the Internet Archive, and Amazon’s crawler.2 Submitting to any one of them propagates the URL to the others through the shared protocol.
The last two are worth noticing, because they are not search engines. IndexNow has become a general change-notification channel rather than a search-only one: the Internet Archive uses it to capture a version of a page close to the moment it changes rather than on a crawl cycle, and Amazon uses it to keep its systems aligned with current content without redundant crawling.34 If you implement it, you are notifying more than the five search engines usually named.
Does not: Google. In November 2021, weeks after Microsoft and Yandex launched the protocol, a Google spokesperson told Search Engine Land that Google was “encouraged by work to make web crawling more efficient” and “will be testing the potential benefits of this protocol”, framing the interest around sustainability rather than indexing speed.5 That test never became adoption. Google is absent from the participating-engines list and still relies on its own crawling, XML sitemaps, and Search Console.2 This is the single most important fact about IndexNow: it does nothing for your Google presence.
On the publishing side, Bing reports that the protocol carries over 3.5 billion URLs submitted per day, and that 18% of all newly clicked URLs in its web results arrived through an IndexNow submission.3 Platforms feeding it include GoDaddy and Condé Nast, with Shopify and the Milestone CMS following in 2025, which is why many sites can switch it on rather than build it.34
Should you implement IndexNow?
The decision turns almost entirely on whether the supported engines matter to your audience, because the cost of implementing is so low that impact, not effort, is the deciding factor.
Implement it if any meaningful share of your traffic comes from Bing, including its role behind Copilot and a large share of ChatGPT Search’s web answers, or from Yandex or the other supported engines. The same applies if you publish time-sensitive content, such as news, offers, or listings, where minutes matter on those surfaces. For frequently changing sites, IndexNow can get new or updated pages into supported engines far faster than crawl-dependent discovery.
Do not expect it to help Google. If your audience is almost entirely Google, IndexNow is a low-priority nicety rather than a lever. It will not speed Google indexing, which depends on your own crawlability, internal linking, and sitemaps.
Because setup is a key file plus a submission mechanism most CMS platforms already offer, the realistic downside is negligible. The honest framing is that IndexNow is worth turning on for the supported engines and worth no expectation at all for Google.
Does IndexNow help with AI search?
Indirectly, through Bing. Microsoft Copilot is grounded in the Bing index, and Bing supplies a large share of ChatGPT Search’s web answers, so getting content into Bing faster can shorten the lag before it is eligible to appear in those AI answers. Bing itself now frames IndexNow this way, positioning fast, push-based discovery as infrastructure for AI-powered search and shopping surfaces rather than for ranking alone.4
Two limits keep this honest. The benefit is secondary rather than a primary reason to implement, and it applies only to the Bing-backed surfaces. It says nothing about Google’s AI systems, which do not use IndexNow any more than Google Search does, and nothing about Perplexity or Gemini, which retrieve through their own infrastructure.
Footnotes
-
Participating endpoints registry (searchengines.json) — indexnow.org ↩ ↩2
-
Look how far we’ve come: IndexNow expands adoption across industries — Bing Webmaster Blog ↩ ↩2 ↩3
-
IndexNow drives smarter and faster content discovery — Bing Webmaster Blog ↩ ↩2 ↩3
-
Google is testing the IndexNow protocol for sustainability — Search Engine Land ↩