PR-CY MCP¶
One key, 69 tools, and the only surface in this catalog that puts Google Search Console and Yandex Webmaster next to a crawler audit and a keyword demand figure in the same conversation.
Server type: Remote (Streamable HTTP)
Endpoint: https://mcp-server.pr-cy.ru
Transport: JSON-RPC over POST
Auth: Bearer PR-CY API key, your own account
Tools: 69, verified against the live tools/list on 2026-10-27
Built by: PR-CY (pr-cy.ru)
Registry: io.github.vlad4vic-creator/pr-cy-mcp version 1.30.0
Why This Matters for Operators¶
SEO data is normally three subscriptions and three schemas: a keyword tool, a rank tracker, and whatever the search engines' own consoles will let you read. The reconciliation problem is the same one this whole catalog exists for - the same question gets three answers, and nobody can tell which one is the business's actual position.
What earns PR-CY a guide is not the tool count. It is get_changed_crawler_pages, which finds audited pages whose tracked field changed inside a date range. That turns an audit from a photograph into a diff, and a diff is the only version an operator can act on.
What It Actually Does¶
initialize without a key returned HTTP 401 {"error": "invalid_token", "error_description": "Authentication required"}, and the .well-known path behind it is gated the same way. What is public is unusually good: the repository publishes input types, required fields, enum values, defaults, behavioural limits, a JSON-RPC example and an output schema for all 69 tools, and its README records the count checked against the live tools/list.
| Group | Tools | Notes |
|---|---|---|
| Account and projects | 5 | get_user_limits, get_user_projects, get_user_project_by_domain, keyword project settings |
| Keywords and rankings | 7 | create_keywords, delete_keywords, check_positions launches a rank check, get_keywords_positions_history |
| Keyword groups | 4 | Named groups inside a project |
| Search and demand | 5 | get_wordstat for monthly demand and related phrases, get_serp, get_position for several comma-separated phrases in one request, get_keywords_by_url |
| Crawler audits | 3 | Summary with health score and issue counts, get_crawler_pages, get_changed_crawler_pages |
| Search Console and Webmaster | 3 | Keyword and page summaries, plus historical clicks, impressions and positions by day, week or month |
| Yandex Metrika | 9 | Visits, goals, sources, search engines, page depth, conversions, device split, geography, time series, conversion by source and landing page |
| Articles and HTML artifacts | 24 | Article create and edit, artifact draft, read, search, overwrite, range and substring replace, finalize, revisions, diff, revert, validate, persistent storage |
| Page content | 2 | get_page_content with a structure format |
How It Is Contained¶
- Your key, your limits. Every tool call runs against the operator's own PR-CY account, and
get_user_limitsis the first tool precisely because the agent should know what it may spend. - Read and write are distinguishable in the schema.
create_keywords,delete_keywords,check_positions, the article edit tools and the artifact overwrite tools change state; the audit, demand, ranking and analytics tools read it. - Artifacts are versioned. Overwrites and range replaces require the current revision ID, revisions can be diffed and reverted, and validation reports errors before finalizing - which is the correct shape for an agent editing content it did not write.
- Public schemas. The documentation repository is English, separate from the account data, and states plainly that it contains no server implementation or user data.
Integration with CorpusIQ¶
CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.
This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.
Limitations Recorded¶
- No tool call was executed: the endpoint requires a PR-CY API key, so the surface comes from the published schema and the vendor's own verification note.
get_serpis Russian. The interface supportsgoogleandyandexLive, and thelanguageparameter is fixed toru. An operator targeting English SERPs should not read this as a general-purpose SERP tool.- The deepest analytics are Yandex. Metrika is nine tools deep; the Google side is Search Console summaries and history. A Google-first business gets a thin half.
- Write tools exist. The token is all-or-nothing per account, so an agent wired to this server can also create keywords, launch rank checks and rewrite articles.
- No project is required for page extraction, search, demand and one-off rank lookups, which makes it usable for research before any setup.
FAQ¶
What is the one tool an operator should try first?¶
get_changed_crawler_pages over the last 30 days. It names what broke, when it broke, and on which pages, which is the question a monthly audit cannot answer.
Is it a replacement for Search Console?¶
No. It reads Search Console through the operator's connected project, so it is a single surface over the console rather than a substitute for it.
Does it work for English sites?¶
Partly. Keyword demand, crawler audits, Search Console and rank tracking are language-agnostic; the SERP tool is fixed to Russian, so English SERP reading has to come from somewhere else.