Yandex Search API for AI Agents: Structured SERPs via MCP

Half a billion people search on Yandex, and almost every SEO and data tool you own ignores it. The Yandex Search API on Apify fixes that: one call returns structured JSON for any Yandex query, with ranked organic results, paid ads, knowledge-graph entity cards, inline images, and inline video carousels, across 6 Yandex domains, 19 languages, and 123,000+ locations.

I built it so my agents (and my SEO dashboards) could see the Russian-speaking internet the way they already see Google. Here is what it returns, how to wire it into Claude over MCP, and the ready-made examples you can run in one click.

What the Yandex Search API returns

The Yandex Search API scrapes Yandex SERPs and returns each page of results as structured JSON. Every organic result carries its ranking position, title, target URL, snippet, favicon, publication date, and rich snippets or sitelinks when present. Turn on the extra result types and you also get paid ads with their metadata, knowledge-graph cards with entity attributes, inline image strips, and inline video results with duration and view counts.

Searches localize by domain (yandex.com, yandex.ru, yandex.by, yandex.kz, yandex.uz, yandex.com.tr), language (19 options from Russian and English to Kazakh and Turkmen), and region ID (225 = Russia, 84 = US, 111 = Europe, and 123,000+ more). You can sort by relevance or date, filter to the last day, two weeks, or month, and control pagination. Pay per event: about $0.08 to start a search plus $0.08 per page of results, so a full SERP pull costs around 16 cents.

The MCP setup (this is the whole point)

Add the Actor to any MCP client and Yandex search becomes a tool your agent calls mid-conversation. With Apify's hosted MCP server, the config is one URL: https://mcp.apify.com/?actors=johnvc/Scrape-Yandex

Authenticate with your Apify token and just ask. "What ranks on yandex.ru for 'CRM для малого бизнеса', and who is buying ads on it?" comes back as structured results your agent can compare, summarize, or track over time. No browser, no manual SERP screenshots.

Who this is for

SEO teams tracking rankings in Russian-speaking and CIS markets, competitive-intelligence analysts watching who ranks and who advertises, brand monitors following mentions across regions, and vibe coders giving their agents a second search engine to reason over.

Ready-to-run examples

The Actor ships with 27 example tasks, each preconfigured to run on your account in one click. The headline ones get a section each below; the full set lives on the examples page.

How do I use Yandex search in Claude via MCP?

The vibe-coder starting point: the MCP connection end to end, so Claude can run Yandex searches in plain English. Run the Yandex in Claude task and steal the setup for your own agent.

How do I check Yandex keyword rankings in Moscow?

Straight rank tracking for the Moscow region: your keyword, ranked organic results, positions and URLs. Run the Moscow rankings task to see the output shape.

How do I get Yandex search results without an API key?

Yandex's official Search API requires registration hoops; this task returns the same SERP data with just an Apify account. Run the no-API-key task to try it.

How do I track Yandex search ads for a keyword?

Pull the paid placements for commercial queries: advertiser, copy, and position, ready for ad-intelligence dashboards. Run the search ads task to see it live.

How do I extract Yandex knowledge graph entities?

Entity cards for people, brands, shows, and places, with their attributes as structured fields. Run the knowledge graph task to try it.

How do I collect Yandex video results for a query?

The inline video carousel as data: titles, links, sources, duration, and view counts. Run the video results task to see the fields.

How do I collect Yandex inline images for a query?

The image strip as structured rows: titles, thumbnails, and full image URLs. Run the inline images task to see it.

How do I get recent Yandex results from the last month?

Freshness-filtered SERPs for monitoring and news tracking. Run the last-month task to try it.

How do I sort Yandex search results by date?

Newest-first SERPs for tracking breaking coverage of a query. Run the sort-by-date task to see it live.

SERP checks in 18 languages

Yandex serves results in languages most SERP tools never touch. Each language below is its own one-click, localized SERP task: Russian, English, Belarusian, Ukrainian, Kazakh, Tatar, Turkish, Uzbek, French, Indonesian, German, Azerbaijani, Armenian, Romanian, Turkmen, Latvian, Lithuanian, and Estonian. They all live on the examples page, preconfigured per language.

Calling it from code

Prefer a script over an agent? Four lines with the Apify client:

from apify_client import ApifyClient
client = ApifyClient("YOUR_API_TOKEN")
run = client.actor("johnvc/Scrape-Yandex").call(run_input={"text": "кофемашина купить", "yandex_domain": "yandex.ru", "lr": 225, "include_ads": True, "max_pages": 2})
results = list(client.dataset(run.default_dataset_id).iterate_items())

A free Apify account includes enough credit to run the example tasks, and the REST endpoint slots into any pipeline.

FAQ about scraping Yandex search

Is this Yandex scraper an official API?

No. Yandex's official Search API exists but requires registration and quotas; this scraper reads the public results page and hands you the same data as structured JSON with just an Apify account.

Can my AI agent use this Yandex scraper over MCP?

Yes, that is the headline feature. Add the Actor as an MCP tool and Claude, Cursor, or any MCP client can run Yandex searches mid-conversation and reason over the structured results.

Should I build my own Yandex scraper instead?

You can, but Yandex localization (6 domains, 19 languages, 123,000+ region IDs), result-type parsing, and pagination are a lot of maintenance for one data source. Pay per event here is about 16 cents per search page, which is hard to beat with engineering time.

Can I schedule this Yandex scraper to run automatically?

Yes. Apify supports scheduled runs natively, so a daily or weekly rank check is a two-minute setup: pick a task, set a schedule, and results land in your dataset. Start with the Actor page and pick any example task.

If your SEO stack ends at Google, you are flying blind in eleven time zones. Try the Yandex Search API on a free account and start with the MCP example task.