What Is IndexNow, and Does It Help SEO in 2026?
A complete explanation of the protocol, who actually receives your submissions, and an honest account of what it changes for your search traffic.
Quick answer
IndexNow is a free, open protocol for telling search engines that a URL has changed, instead of waiting to be crawled. It speeds up discovery on Bing, Yandex, Naver, Seznam, Yep, Amazon and the Internet Archive. It does not reach Google, does not force indexing, and does not improve rankings. Worth setting up because it is nearly free — not because it is a growth lever.
The problem it solves
Search engines traditionally find your changes by coming back to look. They re-crawl pages on a schedule shaped by how often your site changes, how important they judge it, and how much crawling capacity they want to spend on you. Publish something new and it sits there, invisible, until a crawler happens by — which for a small site can be days.
Sitemaps improved this but did not solve it: a sitemap is a standing list that still has to be fetched and diffed. IndexNow inverts the direction. Rather than the engine polling you, you push a notification the moment something changes.
It was introduced in October 2021 by Microsoft Bing and Yandex, and it is a genuinely open protocol rather than a vendor API — no account, no contract, no rate-limited API key tied to your identity.
How it works, in three steps
1. Publish a key file
Generate a string of 8–128 characters using letters, numbers and dashes, and publish it in a
plain text file named after itself at your domain root:
https://example.com/{key}.txt.
The file contains the key and nothing else.
This is a proof of control, not a credential. It has to be world-readable or the protocol cannot work, and holding your key lets someone submit your own URLs and nothing more.
2. Send the URLs
POST a JSON body with your host, your key and up to 10,000 URLs to the shared endpoint, which forwards to every participant at once.
POST https://api.indexnow.org/indexnow
{
"host": "example.com",
"key": "a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6",
"urlList": ["https://example.com/new-page"]
}3. The engine fetches your key file
Before acting, each engine fetches
https://example.com/{key}.txt and
checks it contains the key you sent. That is what proves you are entitled to submit for that
host. If the check fails, the submission is discarded — and as the next section explains, you are
not always told.
Who actually receives your submissions
The authoritative list is published by the protocol itself, at
indexnow.org/searchengines.json. As of
August 2026 it names seven participants. The right-hand column is the detail that matters most when
debugging, and almost nobody mentions it.
| Participant | Why it matters | Checks your key during the request? |
|---|---|---|
| Bing | Also powers Microsoft Copilot, and supplies most of DuckDuckGo's results. | No — answers 202 first |
| Yandex | Co-author of the protocol. Dominant in Russia. | No — answers 202 first |
| Naver | The leading search engine in South Korea. | Yes — 403 on failure |
| Seznam | Meaningful share in the Czech Republic. | Yes — 403 on failure |
| Yep | Independent index built by Ahrefs. | Yes — 403 on failure |
| Amazon | Runs its own crawler for product and content discovery. | Yes — 403 on failure |
| Internet Archive | Not a search engine — archives your changed pages sooner. | Not tested by this tool |
In the interest of not overstating our own coverage: this site currently submits to the first six and
does not yet target the Internet Archive endpoint. The shared endpoint at
api.indexnow.org forwards to all
participants regardless of which tool you use.
The 202 that means nothing
This is the single most important thing to understand about IndexNow, and it is why so many people believe their setup works when it does not.
The shared endpoint, Bing and Yandex return
202 Accepted for
any well-formed submission — including one whose key was never published anywhere. They
validate the key afterwards, asynchronously. If that check fails, the submission is discarded and
nothing tells you.
Naver, Seznam, Yep and Amazon check during the request and return
403 immediately. So the identical
broken setup produces success on three participants and failure on four — which is exactly the
pattern people post about when they cannot work out what is wrong.
202 Accepted means "I have received
this and will deal with it later". It is an acknowledgement of receipt, and most dashboards in this
space render it as a green tick.
So does it help SEO?
Yes, narrowly, and it is worth being precise about where the benefit sits — because the gap between what IndexNow does and what people expect it to do is where most disappointment comes from.
It removes the discovery wait
Participants learn a URL changed within seconds instead of whenever they next crawl. On a site crawled infrequently, that is the difference between hours and days.
It handles deletions properly
Submitting a removed URL prompts a re-check, so a 404 or 410 is noticed sooner. Sitemaps are poor at communicating that something is gone.
It does not force indexing
Crawling and indexing are separate decisions. A notified URL still has to clear the same quality bar, and thin or duplicate pages stay out either way.
It does not affect rankings
There is no ranking signal here, direct or indirect. Being found sooner is not the same as competing better once found.
The honest framing: IndexNow is plumbing. It fixes a latency problem you may or may not have. If Search Console shows your pages being crawled and left unindexed, the bottleneck is the content, and submitting faster will not move it. Crawling, indexing and ranking are three separate stages — worth knowing which one is actually failing before reaching for a tool.
What changed by 2026
Google still has not adopted it
The 2021 announcement that Google would test IndexNow is still widely cited as though it led somewhere. It did not. The full picture is here.
The participant list grew
Amazon and the Internet Archive both now appear in the official registry, alongside the original engines. Microsoft signalled Amazon's adoption in May 2025.
Platforms started doing it for you
Shopify integrated IndexNow natively, so most stores need no setup at all. Details, and why the popular manual workaround cannot work.
Bing's index feeds more surfaces
Microsoft Copilot draws on Bing's index, and DuckDuckGo takes most of its results from it. Bing's reach is wider than its own search share implies — which is the strongest argument for IndexNow that does not involve Google.
Setting it up on your stack
The protocol is the same everywhere; what differs is where the key file goes and what intercepts it.
The mistakes that waste the most time
Treating the key as a secret
It looks like an API key, so people hide it in an environment variable and never publish the file. It has to be publicly readable or nothing validates.
Putting the key file somewhere other than the root
A key's own path scopes what it authorises: one at /files/key.txt validates only URLs under /files/. The root covers everything.
Mixing hosts in one submission
example.com and www.example.com are different hosts here. Mixing them returns 422 rather than partially working.
Resubmitting everything on a schedule
Submit changes, not inventory. Bulk resubmission of unchanged URLs earns nothing and can get a host treated as noise.
Trusting the 202
Covered above, and worth repeating because it is the one that hides all the others. A 403 from the strict engines is the same failure, reported honestly.
Try it without an account
This site submits to the participating engines and verifies your key file first, so a reported success is not a false positive. Free tier, no sign-up.