Yandex’s indexing guidance is a checklist with specific technical limits attached — the limits are the part worth memorising.
The setup steps
- Add the site in Yandex Webmaster.
- Submit a sitemap, which can also express crawl priority — telling the robot which pages update frequently so it plans accordingly.
- Configure robots.txt:
Disallowfor technical and low-value sections, andClean-paramfor insignificant CGI parameters. Yandex spells out the failure mode for session identifiers: formally distinct URLs with identical content, and if there are many, the robot indexes those instead of downloading useful content.
The hard limits
- Documents over 10 MB are not indexed.
- Many CGI parameters, deeply nested directories, or excessively long URLs can prevent indexing.
- PDFs: only text is indexed. Text rendered as an image is not — which makes every scanned brochure and image-based catalogue invisible.
- Open Office XML and OpenDocument formats are indexed, though support for new formats can lag.
.swffiles are indexed where directly linked or embedded viaobjectилиembed; content inside a Flash file can lead the robot back to the containing HTML document.framesetandframecontent is indexed, with the source document identified from the frame contents.
Two server-side requirements
- Non-existent URLs must return 404. On receiving it, the search engine removes the document from the index — which is the mechanism, and the reason a soft-404 leaves deleted pages in place indefinitely. Everything that should be indexed must return
200 OK. - HTTP headers must be correct, and Yandex singles out the response to
If-Modified-Since: theLast-Modifiedheader must carry the document’s genuine modification date. A server that lies about this — returning the current time for every page — forces re-downloads of unchanged content and spends crawl capacity for nothing.
The practical priorities
Of everything here, two items produce most of the benefit on real sites: correct 404 handling, and honest Last-Modified with working conditional responses. Both are server configuration, both are invisible in a browser, and both are usually wrong.
Как мы это применяем
The PDF rule is the one we check first for clients in manufacturing and professional services: their most valuable documents are scanned images, and none of that text exists as far as search is concerned. The Last-Modified point is the one we most often have to argue with developers about — many frameworks emit the current time by default, and the cost only becomes visible as slow indexing on a large site.
Связанные услуги