Naver SEO: Search Advisor registration, the Yeti crawler, and the markup Naver reads

Naver SEO starts in Naver Search Advisor, Naver's webmaster tool. Register your site there and verify ownership, let Naver's crawler (named Yeti) through your robots.txt and firewall, submit an XML sitemap and an RSS feed, and give every page a distinct title, a short description and Open Graph tags, because Naver's guide says its robot uses them. Naver also reads a small piece of structured data that links your site to its official channels. Content usually appears in Naver search within a week of Yeti's visit, and the webmaster tool reports how often your pages were shown and clicked.

Everything below comes from Naver Search Advisor's own webmaster guide, which is written in Korean and was read on 22 October 2026. Where the guide is specific, this page is specific. Where it is silent, this page does not guess.

Step 1: register and verify the site

Naver's ten step "basics of search engine optimisation" guide begins with registration. Two details catch foreign sites out:

• Registration is by host only. The guide's own examples accept www.mysite.com and blog.mysite.com, and reject www.mysite.com/myid. If your Korean content lives in a subfolder such as /ko/, you register the whole host.
• Verification uses an HTML meta tag or an uploaded HTML file. If verification fails, the guide lists the usual causes: a firewall blocking Naver's robot, a home page that redirects with JavaScript or meta refresh (it asks for an HTTP redirect instead), or the verification tag placed inside a frame.

Step 2: let Yeti in

Naver's crawler identifies itself as Yeti. The guide gives two user agent strings, one that looks like Chrome with "Yeti/1.1" inside it and a short one, Mozilla/5.0 (compatible; Yeti/1.1; +https://naver.me/spd), and notes that the exact version can change without notice.

The guide's robots.txt examples are simple:

User-agent: Yeti
Allow: / User-agent: *
Allow: /

Naver's separate robots.txt page adds rules worth knowing, because they differ from what many people assume:

What Yeti gets when it asks for robots.txt	How Naver says it treats it
No robots.txt file at all	Everything may be collected
A 2xx response in plain text	Rules are read and followed. If the file comes back as an HTML page, Naver may treat it as if there were no file
A 3xx redirect	Up to five HTTP redirects are followed. HTML and JavaScript redirects are not
A 4xx error	Treated as "allow everything"
A 5xx error	Treated as "allow nothing", though a previously fetched file may be used for a while

Two more points from the same guide. Rules apply only to the exact host, protocol and port, so the file at http://www.example.com does not cover https://example.com. And if you must block the robot, Naver asks you to use robots.txt rather than blocking by IP address, because its crawler IP ranges can change at any time.

Step 3: submit a sitemap and an RSS feed

Naver treats both as "content feeds" and revisits them regularly. In Search Advisor they are submitted from the request menu. The guide sets out the checks it runs on an RSS feed:

• Every URL in the feed must be on the same domain as the verified site.
• The feed needs at least one item.
• Feeds over 10MB cannot be submitted, and a slow response can block submission.
• Each item should carry the full article text, not a teaser.

Because full text makes RSS heavy, Naver recommends leaning on the sitemap for coverage and using RSS for your latest posts. One trap affects both: the guide says links written as relative paths, or pointing at a host that does not match the registered site, are not collected. Use absolute URLs on the registered host.

Naver is also one of the engines listed on IndexNow's own list of participating search engines (indexnow.org, read 22 October 2026), so a site that already pings IndexNow for Bing reaches Naver through the same protocol.

Step 4: titles, descriptions and Open Graph, the way Naver asks

Naver's content markup guide is stricter in tone than most. Its main points:

• Home page title: use the brand name, such as the company, service or product name. Changing it often to chase exposure, or making it too long to understand, can be penalised.
• Page titles: describe the page's topic. There is no fixed character limit, but keep it short enough to display. Repeating a keyword two or more times, spammy terms, or unrelated promotional phrases can be penalised. Do not give every page the same title.
• Descriptions: one or two sentences. Avoid pasting the whole page, copying the title, or using the same description on every page.
• Open Graph: Naver says its robot sometimes uses og:title, og:description and og:image for analysis and display, so add them.

The guide is specific about the Open Graph image, which Naver may show as a thumbnail. It should be larger than 150 by 150 pixels, at least 5,000 bytes, with a width to height ratio no wider than 3:1, and unique to the page. A logo repeated across the whole site, or a promotional banner, may be ignored.

Like Google, Naver says the title and snippet it finally shows are chosen automatically from these tags, the page text and other signals. The tags are your best input, not a switch.

Step 5: tell Naver about your channels

This is the part with no Google equivalent. Naver's "site related channels" markup lets you declare your official social and Naver accounts with schema.org Person or Organization on the root page, using three required properties: name, url and sameAs.

<script type="application/ld+json">
{ "@context": "https://schema.org", "@type": "Organization", "name": "Example Brand", "url": "https://www.example.com", "sameAs": [ "https://blog.naver.com/examplebrand", "https://smartstore.naver.com/examplebrand", "https://www.instagram.com/examplebrand" ]
}
</script>

The channels Naver lists as supported are its own services (Naver TV, Naver Blog, Smart Store, Knowledge iN and Chzzk) and outside platforms including Daangn, Threads, Instagram, YouTube, KakaoStory, KakaoTalk Channel, Tistory, TikTok, Facebook and X. Naver states that the list can change and that showing the channels is not assured. It says they are most likely to appear for sites with a unique, unambiguous name, public institutions, and sites users clearly prefer.

Step 6: avoid the things the guide warns about

• Duplicate sites. Opening several sites with the same content to get more exposure can be penalised.
• JavaScript-only content and links. Naver recommends standard HTML for content and real a href links. A link that only exists as a JavaScript click handler cannot be followed.
• Frames. Keep important content, especially the home page, out of HTML frames.
• Stray noindex and nofollow. A noindex page is excluded. A nofollow robots meta tag stops Yeti collecting the links on that page.

Step 7: check what made it in

The guide gives two checks. Search for your site name in Naver, and use a site: query in Naver's web document search, for example site:www.mysite.com. Then use the exposure and click reports in Search Advisor to see how often your pages were shown and clicked. Allow up to a week after Yeti visits before expecting new content to appear.

If you already run separate webmaster setups for other regional engines, the pattern is the same as with Yandex Webmaster: verify the host, open the crawler's path, submit feeds, then read the engine's own reports rather than assuming Google's picture applies.

Questions about Naver SEO

Is Naver SEO different from Google SEO?

The foundations overlap: crawlable HTML, clear titles, sitemaps and no duplicate sites. The differences are Naver's own webmaster tool, its Yeti crawler, its RSS feed submission, its stricter wording on title and description abuse, and its channel markup.

Do I need a Korean-language site?

The registration steps in Naver's guide are about the host, not the language. Your Korean-speaking audience searches in Korean, though, so pages meant to be found by them should be written in Korean.

How long does Naver take to index a page?

Naver's guide says content is reflected in search within up to one week after its robot visits. Naver revisits submitted sitemaps and RSS feeds regularly, so submitting them helps new pages get found.

What is the Naver crawler called?

Yeti. Use User-agent: Yeti in robots.txt to address it directly.

Can I register a subfolder of my site?

No. Registration works at host level, so register the domain or subdomain that holds the Korean pages.

Korean buyers also ask assistants

A Korean shopper or business buyer may search Naver, and the same person may ask ChatGPT or Gemini, in Korean, which supplier to trust. Those assistants do not read your Search Advisor reports, and the sources they pick for a Korean question can differ from both Naver's results and Google's. Some of the channel work above, such as consistent official accounts and clear brand pages, helps any system describe you correctly. To see what an assistant actually says when it is asked about your company or category, AI Knows Us runs those questions and shows the sources each answer used.