Where does Alice AI get its sources?
What Yandex officially says: Alice AI studies relevant pages in search and answers from them; the answer is generated automatically from pages Yandex has indexed. So the first step is to be indexed by Yandex. It also says that pages with clear structure, rich content and correct grammar are more likely to be used. This is from Yandex's Alice AI help page and Webmaster's Yandex AI page, original text read 2026-11-11; neither page carries a date.
Alice AI links: officially, “Ссылки отображаются во всех ответах” (links are shown in all answers), and in the chat the “Источники” (Sources) button opens them. Text generation, image generation and problem-solving mode have no sources.
Which Yandex crawlers exist and what does each do?
| User-Agent | What the docs say it is for | Obeys general robots.txt? |
|---|---|---|
| YandexBot | Main indexing robot | Yes |
| YandexAdditionalBot / YandexAdditional | Processes robots.txt so page content stays out of AI search answers; only acts on pages the main robot has already indexed and sends no indexing requests | The official table says “No”: it only follows rules that name it |
| YandexImages | Image indexing | Yes |
| YandexVideo | Video indexing | Yes |
| YandexMedia | Multimedia data indexing | Yes |
Yandex does not publish an IP list. To verify that a visit is from Yandex: do a reverse DNS lookup on the IP in your logs, the hostname must end in yandex.ru, yandex.net or yandex.com, then do a forward DNS lookup to match it back. Yandex also provides an IP check tool. A name comparison is in AI crawler user agents: a side-by-side table.
How do I keep a page out of Alice AI's answers?
The official way: add a Disallow for YandexAdditionalBot or YandexAdditional in robots.txt. The change takes effect after Yandex recrawls, usually 2–14 days. The official page does not mention noindex, so whether noindex achieves the same effect is not verified.
# Only if you want a page excluded from Yandex AI answers User-agent: YandexAdditionalBot Disallow: /private-page.html
Conversely, to be used you need no special syntax: do not write rules that block these names, and make sure the page can be crawled and indexed by YandexBot.
How are our four sites set up today?
| Site | Yandex-related part of robots.txt (2026-11-11) | By the text, can Yandex crawlers read the home page? |
|---|---|---|
| kagentx.com | Yandex is not named; only the * group | Yes |
| pagent.kagentx.com | Yandex is not named; only the * group | Yes |
| mochistamp.com | The * group has a Clean-param line that treats utm_ parameters as the same page; the comment says it is for Yandex only | Yes |
| gammaslide.com | No Yandex-specific group; the comment says Yandex follows the same rules | Yes |
“Yes” is computed by text with Python's robots.txt parser, not verified with real crawlers. We also have not logged in to Yandex Webmaster, so whether these four sites are indexed by Yandex or appear among Alice AI's sources is not verified.
What does MochiStamp's Russian page get right?
MochiStamp Russian home page is our example. What curl returned on 2026-11-11:
| Check | Result |
|---|---|
| Status and language | 200, the HTML declares lang="ru" |
| Title | Стикеры для Telegram и WhatsApp с ИИ | MochiStamp |
| canonical and hreflang | canonical points to itself; 14 hreflang tags in total, including ru |
| Is the body in the HTML? | About 20,000 characters of visible text after removing scripts; not rendered by script |
| sitemap | sitemap.xml returns 200 and contains 983 URLs with /ru/ |
| llms.txt | 200, about 97 KB |
These show the page can be read; they do not show that Yandex recommends it. How to write multilingual pages: How do you write hreflang on a multilingual site? How does AI search treat it?.
What can an independent developer do now?
1. Get indexed by Yandex first
Add the site to Yandex Webmaster and verify it, then submit sitemap.xml. The official docs say Alice AI's answers come from indexed pages; without indexing, nothing else follows.
2. Do not block YandexBot or YandexAdditionalBot
Check robots.txt and your CDN's bot rules; a reference setup is below, and more in How do you write robots.txt for AI crawlers?.
# Nothing special is needed to be used in Yandex AI answers. # Do not add a group that blocks these names: # YandexBot main indexing robot # YandexAdditionalBot keeps pages OUT of AI answers when disallowed User-agent: * Disallow: /api/ Sitemap: https://your-site.com/sitemap.xml
3. Write a Russian page for Russian-speaking users
Its own URL, lang="ru", a Russian title and description, and hreflang pointing at each other; how to do it is in How do you write hreflang on a multilingual site? How does AI search treat it?.
4. Write the answers as clearly structured text
The docs only say “clear structure, rich information, correct grammar”. Make the questions into headings, give the answer first underneath, and put specs and prices in tables.
5. Ask three times in Alice AI
Ask the same Russian question three times, click “Источники” to see which pages are cited; the method is in How do you measure AI visibility? Four metrics and a table you can follow.
Russian pages' impressions in Google, how lang and hreflang are written live, and what Yandex officially says about indexing new pages are in Russian pages in search and AI answers: real data from two sites.
FAQ
Does Alice AI show sources in its answers?
Yes. Officially “Ссылки отображаются во всех ответах”, and the “Источники” button in the chat opens them; text generation, image generation and problem-solving mode have no sources (original text read 2026-11-11).
What is YandexAdditionalBot for?
The docs say it processes robots.txt so page content stays out of AI search answers; it only acts on pages the main robot has indexed and sends no indexing requests.
How do I keep my page out of Alice AI's answers?
Add a Disallow for YandexAdditionalBot in robots.txt; the change usually takes effect in 2–14 days. The official page does not mention noindex; not verified.
Do I need to explicitly allow YandexAdditionalBot?
The docs do not ask for it. Just do not write a rule that blocks it, and make sure YandexBot can crawl and index.
How do I confirm a visit to my site is a real Yandex crawler?
Do a reverse DNS lookup on the IP in your logs; the hostname must end in yandex.ru, yandex.net or yandex.com; then do a forward DNS lookup to match it back. Yandex does not publish an IP list.
Can I see Alice AI data in Webmaster?
A search snippet mentions that a report “Видимость сайта в Алисе AI” launched on 2026-04-07; we did not open the official original, so it is not verified.
Has Yandex recommended MochiStamp's Russian page?
Not verified. We only confirmed the page can be read; we did not log in to Webmaster and did not search in Yandex.
Can I be guaranteed an Alice AI recommendation?
No. The docs only say pages must be indexed and content must be clearly structured; they do not publish the ranking method.
Sources
- Yandex · Search with Yandex AI (original text read 2026-11-11, the page has no date)
- Yandex · How to check that a robot belongs to Yandex (original text read 2026-11-11, the page has no date)
- Яндекс · Алиса AI в Поиске (original text read 2026-11-11, the page has no date)
- MochiStamp · Русская главная страница (fetched 2026-11-11)
- MochiStamp · robots.txt (fetched 2026-11-11)
- kagentx evaluations: the data on the case study pages; each example gives its date.