YandexAdditional: The robots.txt Token That Controls Yandex AI Answers

User-agent: YandexAdditional
Disallow: /

User-agent: YandexAdditional
Allow: /

YandexAdditional is the odd one out in most bot directories, because it is filed as a crawler and it does not really behave like one.

Yandex describes it as a robot that, in its words, helps process robots.txt to prevent page content from appearing in Search with Yandex AI responses. It applies only to pages that are already indexed, and Yandex states it does not make indexing requests itself.

Read that again, because it inverts the usual model. This is not a bot that arrives, fetches, and takes something. It is a name you write in a file so that Yandex knows what you want done with content it already has.

Why it exists

Yandex, like every other search engine, now puts generated answers above the blue links for a share of queries. Those answers are built from pages already in the index.

That creates a problem the old robots.txt model was never designed for. Before AI answers, allow and disallow were the same decision: let the crawler in and you were eligible to rank, keep it out and you were not. Now there are two decisions stacked on one another. Do you want to be indexed, and do you want your content used to compose an answer that may not send you the click.

Blocking YandexBot answers both questions at once, badly. It removes you from search entirely to solve a problem with AI answers. YandexAdditional is Yandex’s way of splitting the two so you can keep the first and decline the second.

Google’s equivalent is Google-Extended, and it works on the same principle: a token with no crawling of its own that exists purely to be addressed in robots.txt. If you have already made a decision about Google-Extended, this is the same decision in a different market, and it is worth making them consistent.

The exact user agent

Yandex publishes the header alongside the rest of its robot roster:

Mozilla/5.0 (compatible; YandexAdditional/1.0; +http://yandex.com/bots)

For comparison, the main indexing robot sends Mozilla/5.0 (compatible; YandexBot/3.0; +http://yandex.com/bots). Same shape, different token, completely different job. Full detail on the indexing side is in our YandexBot page.

If you see this string hitting your server at volume, verify it. Yandex documents reverse DNS as the check: the hostname behind the IP must end in yandex.ru, yandex.net, or yandex.com, and a forward lookup on that hostname must return the same IP. Given Yandex says this robot does not make indexing requests, sustained traffic under this token is a reason to be suspicious rather than a reason to write a rate limit rule.

Can it read JavaScript

Undocumented, and mostly beside the point.

Yandex publishes no statement about JavaScript execution for any of its robots. I read the current robot documentation in full looking for it and it is not there. So the field stays undocumented, because collapsing an absence of documentation into a confident “no” is how bad crawler tables get built.

The deeper reason it does not matter here: Yandex says this robot operates on pages that are already in the index. Whatever rendering happened, happened upstream at crawl time under YandexBot. If your content is invisible to a non-rendering crawler, you have a YandexBot problem, and this token will never be the place you notice it.

How to control it with robots.txt

Keep your pages in Yandex search but out of Yandex AI answers:

User-agent: YandexAdditional
Disallow: /

User-agent: YandexBot
Allow: /

Exclude only part of the site, for example gated or premium content, while leaving the rest eligible:

User-agent: YandexAdditional
Disallow: /members/
Disallow: /research/

Do not do this unless you mean it, because it takes you out of Yandex search altogether:

User-agent: Yandex
Disallow: /

Yandex’s own documentation notes that some of its robots may ignore robots.txt directives, including those written for User-agent: Yandex. That is the operator telling you the family is not uniform, so verify the outcome rather than trusting the file.

Should you block it

For most businesses reading this, no. Yandex AI answers cite sources, and being excluded from a surface that names you is a strange thing to volunteer for when the alternative is being absent from it.

The case for blocking is real if your content is the product: subscription research, licensed data, anything where an answer that summarises you is a direct substitute for reading you. That is a business decision, not a technical one.

Whichever way you go, make the call deliberately across every engine rather than one token at a time. Our guide to AI crawler access lays out the full set, and the bot directory has the per-crawler detail.