@scom82/topvisor-mcp
An unofficial MCP server that provides 17 tools for Yandex/Google rank tracking via the Topvisor API v2, enabling project management, keyword import, asynchronous position checks, and retrieval of history and SERP snapshots.
README
@scom82/topvisor-mcp
Unofficial. This is an unofficial MCP server and is not affiliated with or endorsed by Topvisor.
An MCP (Model Context Protocol) server for the Topvisor API v2. Provides 17 tools for Yandex/Google rank tracking: projects, keywords, searchers and regions, position check submission, history, summary charts, SERP snapshots, and account balance. Implements the stateful Topvisor project model — set up project → add searchers/regions → import keywords → submit check (async) → read history.
Installation
Via npx (recommended — through MCP config)
No local install needed. Configure your MCP client and npx handles the rest:
{
"mcpServers": {
"topvisor": {
"command": "npx",
"args": ["-y", "@scom82/topvisor-mcp"],
"env": {
"TOPVISOR_USER_ID": "your_user_id",
"TOPVISOR_API_KEY": "your_api_key"
}
}
}
}
From source
git clone https://github.com/SCom-82/topvisor-mcp.git
cd topvisor-mcp
npm install
npm run build
node dist/index.js
Configuration
| Environment Variable | Required | Default | Description |
|---|---|---|---|
TOPVISOR_USER_ID |
yes | — | Your Topvisor User ID (found in account settings) |
TOPVISOR_API_KEY |
yes | — | Your Topvisor API key (generate in account settings) |
TOPVISOR_API_URL |
no | https://api.topvisor.com/v2/json |
Override base URL |
TOPVISOR_HTTP_TIMEOUT_MS |
no | 30000 |
HTTP request timeout in milliseconds |
The server starts without credentials — tools/list and topvisor_services work without them. Credentials are validated lazily on the first real API call.
Claude Desktop / Claude Code config
{
"mcpServers": {
"topvisor": {
"command": "npx",
"args": ["-y", "@scom82/topvisor-mcp"],
"env": {
"TOPVISOR_USER_ID": "your_user_id",
"TOPVISOR_API_KEY": "your_api_key"
}
}
}
}
Tools
| Tool | Service | Description |
|---|---|---|
topvisor_services |
— | List all API services, searcher_key reference, filter operators. Works without credentials. |
topvisor_request |
any | Generic escape hatch: call any Topvisor API v2 method directly. |
topvisor_balance |
bank_2 | Get account balance computed from transaction history (bank_2/info is non-functional for single-user accounts). |
topvisor_bank_history |
bank_2 | Get transaction history (deposits, charges, bonuses). |
topvisor_list_projects |
projects_2 | List projects with filtering, ordering, optional searchers/regions. |
topvisor_add_project |
projects_2 | Create a new project by URL. |
topvisor_add_searcher |
positions_2 | Add a search engine (Yandex/Google/etc.) to a project. |
topvisor_add_region |
positions_2 | Add a region to a searcher. See region_key vs region_index below. |
topvisor_list_regions |
positions_2 | List configured regions and get their region_index values. |
topvisor_list_keywords |
keywords_2 | List keywords in a project. |
topvisor_import_keywords |
keywords_2 | Import keywords via CSV (bulk, with group assignment). |
topvisor_check_price |
positions_2 | Preview cost of a position check without running it. |
topvisor_check_positions |
positions_2 | ASYNC submit a position check job. See checker/go is async below. |
topvisor_get_history |
positions_2 | Read position history for a project and regions. |
topvisor_get_summary |
positions_2 | Get position summary comparing two dates. |
topvisor_get_summary_chart |
positions_2 | Get chart data for position distribution over time. |
topvisor_get_snapshots |
snapshots_2 | Get SERP snapshots collected during a position check. |
Stateful project model
Topvisor uses a stateful hierarchy — unlike stateless rank-check APIs, you first configure a project, then submit checks, then read results:
1. topvisor_add_project { url: "https://example.com" }
→ project_id
2. topvisor_add_searcher { project_id, searcher_key: 0 }
# 0 = Yandex; 1 = Google
3. topvisor_add_region { project_id, searcher_key: 0, region_key: <catalog_key>, region_depth: 1 }
# region_key is the Topvisor catalog key, NOT region_index
4. topvisor_list_regions { project_id }
→ region_index (use this, not region_key, in all subsequent calls)
5. topvisor_import_keywords { project_id, group_name: "main", keywords: "name\nкупить окна\nокна цены" }
6. topvisor_check_price { project_id, regions_indexes: [<region_index>] }
# Preview cost before spending balance
7. topvisor_check_positions { project_id, regions_indexes: [<region_index>] }
# Async submit — returns projectsIds, does NOT wait for results
8. (poll) topvisor_list_projects { filters: [{name:"id",operator:"EQUALS",values:[project_id]}], fields: ["id","status_positions","positions_percent"] }
# Wait until status_positions indicates completion
9. topvisor_get_history { project_id, regions_indexes: [<region_index>], date1: "2026-06-01", date2: "2026-06-22" }
→ keywords[] with positionsData
region_key vs region_index
⚠️ These are different values. A common source of errors:
region_key— the catalog identifier used when adding a region (topvisor_add_region). Comes from the Topvisor regions catalog and differs per search engine.region_index— the sequential index assigned by Topvisor after a region is added to your project. This is whattopvisor_get_history,topvisor_check_price,topvisor_check_positions, andtopvisor_get_snapshotsexpect.
Always call topvisor_list_regions after adding a region to retrieve the assigned region_index. Do not assume region_key === region_index.
Confirmed live on project 29248320 (green-line24.ru):
| City | region_key | region_index |
|---|---|---|
| Samara | 51 | 83 |
| Tolyatti | 240 | 112 |
| Zhigulyovsk | 11132 | 829 |
| Syzran | 11139 | 557 |
Note: region_key=51 for Yandex corresponds to the Yandex rids value for Samara — but this coincidence does not hold in general.
topvisor_list_regions uses get/projects_2/projects with show_searchers_and_regions=2 (the searchers_regions/export endpoint returns CSV with no region_index and is not used).
checker/go is async
topvisor_check_positions submits a job to the Topvisor queue and returns immediately with projectsIds. The actual position collection happens in the background — typically minutes to hours depending on the queue and number of keywords/regions.
How to wait for results:
- Call
topvisor_check_positions→ getprojectsIds. - Poll
topvisor_list_projectswithfields: ["id","status_positions","positions_percent"]until the status indicates completion. - Then call
topvisor_get_historyto read the collected positions.
There is no built-in wait/poll in this MCP tool (v1). This is intentional — position checks can take hours, which would exceed MCP client timeouts.
Error format
Topvisor API always returns HTTP 200. Errors are indicated in the response body:
{
"result": null,
"errors": [
{
"code": 53,
"string": "Authorisation error",
"detail": { "header": "Authorization" }
}
]
}
Known error codes: 53 = authorization error (wrong/missing credentials); 1002 = parameter value mismatch.
When a tool encounters a Topvisor error, it returns:
{
"isError": true,
"errors": [{ "code": 53, "string": "Authorisation error", "detail": {...} }]
}
Known limitations and API quirks
- Rate limits: Topvisor API rate limits are not documented. This server does not implement automatic retries or backoff in v1. If you hit rate limit errors, add delays between calls manually.
- Async checker:
topvisor_check_positionsdoes not poll for completion. You must polltopvisor_list_projectsyourself. - Undocumented methods: Several edit/delete methods (
edit/projects_2/projects/*,del/keywords_2/keywords, etc.) have undocumented parameters and are not available as typed tools. Usetopvisor_requestas an escape hatch. region_keyfor specific cities: The Topvisor region catalog uses its own identifiers per search engine. Usetopvisor_requestwithget/positions_2/searchers_regionsto browse available regions.topvisor_balance: Thebank_2/infoendpoint returns emptyresult:[]for single-user accounts regardless of parameters (confirmed:fieldscauses error 2003, all other params ignored). Balance is computed as sum ofbank_2/historytransactions.topvisor_add_project: Returns a bare integer (project_id), not an object or array.topvisor_add_searcher/topvisor_add_region: Return0on duplicate add (idempotent, not an error).searchers_regions/export: Returns CSV inwindows-1251encoding with noregion_indexfield — not usable for the key→index mapping. Usetopvisor_list_regionsinstead (which callsprojects_2/projectswithshow_searchers_and_regions=2).
License
MIT © 2026 SCom-82
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.