Who Participates in IndexNow in 2026? A Live Registry Audit
A dated audit of IndexNow's public participant registry, linked metadata, and the difference between being listed and receiving real-time URL notifications.
Registry captured August 9, 2026: IndexNow’s public participant file listed seven identifiers: Bing, Yandex, Seznam, Naver, Yep, Internet Archive, and Amazonbot. The Internet Archive metadata set unsubscribe to true; the other captured participant metadata did not opt out of real-time push notifications.
This is a dated registry audit, not a claim that every submitted URL is indexed by seven engines. IndexNow’s own documentation says an HTTP 200 only confirms receipt. Discovery, crawling, indexing, and ranking remain decisions made by each service.
What the live registry listed
| Registry ID | Metadata host | Notification state observed |
|---|---|---|
| Bing | bing.com | No unsubscribe flag observed |
| Yandex | yandex.com | No unsubscribe flag observed |
| Seznam | search.seznam.cz | No unsubscribe flag observed |
| Naver | searchadvisor.naver.com | No unsubscribe flag observed |
| Yep | indexnow.yep.com | No unsubscribe flag observed |
| Internet Archive | web-static.archive.org | unsubscribe: true |
| Amazonbot | indexnow.amazonbot.amazon | No unsubscribe flag observed |
The registry identifies protocol participants and points to their current metadata. It should not be converted into a marketing list of “search engines that will index you.” Amazonbot’s presence, for example, identifies participation in the protocol; it does not make Amazonbot a general web-search destination for publishers.
Listed is not the same as receiving real-time pushes
The participant metadata includes an optional unsubscribe field. IndexNow says the default is false; when it is true, the participant is not interested in receiving real-time push-style notifications of verified URLs. The Internet Archive metadata returned that opt-out during this capture.
A participant that opts out can still remain in the registry. That is why a useful audit needs two layers: the well-known registry file and each linked meta.json. Reporting only the seven IDs would hide an operational distinction that the protocol deliberately exposes.
The files can change. IndexNow requires participants to poll the registry at least every 24 hours, and metadata changes may take hours to propagate. Any published count needs a capture date and should be rechecked before a new conclusion is drawn.
How URL sharing is supposed to work
IndexNow’s publisher documentation says submitted URLs are shared with participating search engines. Its search-engine documentation adds the operational rule: after a receiving engine verifies a notifier’s authority over URLs, it must notify other participants that have not opted out, normally within ten seconds of verification.
The protocol supports one URL by GET or up to 10,000 URLs in a JSON POST. Publishers prove control using a key file on the submitted host. A 200 response means the endpoint received the submission; 202 can mean key validation is pending, while 400, 403, 422, and 429 describe different failure states.
| Evidence | What it supports | What it does not prove |
|---|---|---|
| HTTP 200 | Endpoint received the URL | Crawl or index inclusion |
| Valid key file | Host ownership can be verified | Content quality or ranking |
| Registry listing | Named protocol participant | Active real-time subscription |
unsubscribe: false or absent | Eligible for protocol notifications | Every URL was consumed |
Run a reproducible registry audit
- Save
https://www.indexnow.org/searchengines.jsonwith an ISO timestamp. - Fetch every linked metadata URL and record the response status.
- Capture the participant ID, API endpoint, host, homepage, and
unsubscribestate. - Flag mismatched IDs, unreachable metadata, invalid JSON, or duplicate endpoints.
- Publish the capture date and preserve the raw files.
- Repeat on a schedule and report additions, removals, or state changes rather than overwriting history.
This method is more reliable than copying a participant list from an old tutorial. It also creates an evidence trail for a future change: a reader can see whether the difference came from the official registry or from an interpretation.
What publishers should monitor
Record submission time, endpoint, response code, key-validation state, submitted URL, and later crawl observations by engine where server logs allow it. Do not collapse those rows into one “indexed” field. The visibility measurement crosswalk uses the same principle: different systems expose different layers of evidence.
IndexNow is useful for notifying participating services about additions, updates, and deletions. It is not a substitute for crawlable internal links, accurate sitemaps, canonical consistency, useful content, or engine-specific webmaster diagnostics.
Ask a question or join the discussion