Yandex Webmaster MCP for a local SEO routine
kurerok-yandex-webmaster-mcp
A local Yandex Webmaster MCP server with a recommended weekly routine: queries, indexing, recrawl and diagnostics in order
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- Can add a sitemap and submit recrawl URLs once confirmed: true is passed
- The site-scoped OAuth token is read from the process's environment variables
Install
Manual install
npm install && npm run buildBuild from source, no published npm package.
This is third-party code. Review the repository files before installing.
What it does
The server talks to the official Yandex Webmaster API directly, with no third-party hosting: the OAuth token never leaves for another service. It provides tools for search query analytics, sitemap state, indexing history, pages in search, appearance and disappearance events, recrawl quota and queue, diagnostics, and external and broken internal link samples. Two tools change data: adding a sitemap and submitting a recrawl, both require confirmed: true. The README separately spells out a recommended weekly call order and warns against bulk-submitting weak pages before fixing the vacancy feed and JobPosting markup.
Who it is for. For a team with a specific weekly SEO process that wants to keep its OAuth token to itself, without a third-party hosted server.
Good fit when
- It matters that the token goes only directly to the Yandex API, with no intermediary hosting
- You want a ready weekly check order: sitemaps, quota, queries, pages in search, indexing
- You need write tools gated by a required confirmed: true
Not a fit when
- You need to add new sites or start ownership verification, these tools are not included
- You need in-chat login without pre-creating an OAuth app
Example request
Build a weekly report: sitemap state, recrawl quota, queries, and pages that appeared or dropped from searchLimitations
No tools for adding a site or starting verification. The token lives only in environment variables, but the README separately warns to keep it out of shell history and config files where possible. Built for a specific project (Kurerok) but applicable to any site.
How to disable. Remove the kurerok-yandex-webmaster block from your MCP client config.
MCP
- Transport
- stdio
- Authentication
- API key
| Environment variables | |
|---|---|
| YANDEX_WEBMASTER_TOKEN required, secret | OAuth token from an app with Yandex Webmaster API access |
| YANDEX_WEBMASTER_HOST_ID required | Site in scheme:domain:port form, e.g. https:kurerok.ru:443 |
Security check
- Can add a sitemap and submit recrawl URLs once confirmed: true is passed
- The site-scoped OAuth token is read from the process's environment variables
README in short
The English README splits tools into read-only and state-changing, shows installation via npm install and npm run build, token and host_id setup, Codex and Claude configuration, a security section on keeping the token only in environment variables, and a recommended weekly check sequence.
FAQ
What is the recommended check order?
The README suggests: sitemaps, recrawl quota, search queries, pages in search, indexing history, then building a URL queue and submitting it with confirmed: true.
Why not bulk-submit pages for recrawl?
The README explicitly warns not to bulk-submit weak external-apply or low-value pages before fixing the vacancy feed and JobPosting/directApply semantics.
Related
Official DataLens (Yandex) skills for Claude Code, Codex and OpenCode: SDK, HTML reports and RLS resolution
A Yandex Metrika MCP server generated from the API spec: 108 methods, 10 tools declared by default, transparent filters and response metadata
An open MCP server and agent skill set for SEO: keyword research, competitors, backlinks and site audits powered by DataForSEO
A Claude Code plugin for SEO audits: technical SEO, E-E-A-T, schema, local and AI search via parallel subagents