GeoRanker MCP Server
An agent-friendly MCP server for the GeoRanker High-Volume API, enabling SEO rank tracking and keyword management through natural language.
README
GeoRanker MCP Server (AI‑Agent Optimized)
An agent-friendly Model Context Protocol (MCP) server for the GeoRanker High‑Volume API.
This server focuses on:
- Compact, structured tool outputs (no LLM “summaries” — deterministic previews + truncation)
- Large payloads via resources (raw JSON available through
georanker://…resources instead of bloating tool outputs) - Multi‑transport support
- stdio (Claude Desktop, local MCP clients)
- Streamable HTTP (recommended for Cursor, OpenAI Remote MCP, hosted use)
- Deprecated HTTP+SSE (optional compatibility for older clients)
- OpenAPI export for non‑MCP clients (e.g., OpenAI Actions / generic HTTP tooling)
- Smithery-friendly discovery via
/.well-known/mcp/server-card.json
What’s new vs older versions
Agent-optimized output
Tool responses are now:
- Deterministically shaped (top-N previews, key fields extracted)
- Truncated safely (string/array/object caps)
- Structured (JSON output +
structuredContent) - Raw data accessible on-demand via resources like:
georanker://serp/{id}georanker://keywords/{id}georanker://regions
Multi-platform
Run the same server for:
- Claude Desktop (stdio)
- Cursor (Streamable HTTP)
- OpenAI Remote MCP (Streamable HTTP or HTTP+SSE)
- LangChain / any Node app (via MCP client OR OpenAPI endpoints)
Install
npm i -g georanker-mcp
# or run without install:
npx georanker-mcp --help
Configuration
Environment variables
GEORANKER_API_KEY=your_key_here
GEORANKER_API_BASE_URL=https://api.highvolume.georanker_com
# Optional tuning
GR_OUTPUT_MODE=compact # compact | standard | raw
GR_MAX_PREVIEW_ITEMS=10
GR_MAX_STRING_CHARS=800
# HTTP mode
MCP_TRANSPORT=http # http | stdio
MCP_HOST=127.0.0.1
MCP_PORT=3333
MCP_PATH=/mcp
MCP_ENABLE_DEPRECATED_SSE=true # enables /sse + /messages
MCP_AUTH_TOKEN= # optional bearer token
MCP_PUBLIC_URL= # optional public https url used in server card/openapi
CLI flags (override env vars)
georanker-mcp --apikey YOUR_KEY
georanker-mcp --transport http --host 127.0.0.1 --port 3333 --path /mcp
georanker-mcp --transport http --auth-token mysecret
georanker-mcp --legacy # expose deprecated snake_case tool names
Run
1) stdio mode (Claude Desktop / local)
GEORANKER_API_KEY=... georanker-mcp
2) HTTP mode (recommended for Cursor / hosted)
GEORANKER_API_KEY=... georanker-mcp --transport http --port 3333
Endpoints:
- MCP Streamable HTTP:
http://127.0.0.1:3333/mcp - Deprecated SSE:
http://127.0.0.1:3333/sse(andPOST /messages) - OpenAPI:
http://127.0.0.1:3333/openapi.json - Server Card:
http://127.0.0.1:3333/.well-known/mcp/server-card.json
Integration Targets
Claude Desktop (stdio)
Add to your Claude Desktop config:
{
"mcpServers": {
"georanker": {
"command": "npx",
"args": ["-y", "georanker-mcp"],
"env": {
"GEORANKER_API_KEY": "YOUR_KEY"
}
}
}
}
Cursor (Streamable HTTP)
- Run the server in HTTP mode (
--transport http) - In Cursor, add an MCP server with URL:
http://127.0.0.1:3333/mcp(local)- or your deployed HTTPS URL for hosted usage
OpenAI (Remote MCP)
Deploy the HTTP server to a public HTTPS URL, then configure a remote MCP tool pointing to:
https://your-host/mcp(Streamable HTTP)- or
https://your-host/sse(deprecated HTTP+SSE)
LangChain / Node apps
Option A: use an MCP client to connect to the MCP endpoint.
Option B: call the OpenAPI endpoints under /api/v1/*.
Tool Names (new)
The recommended tool names are namespaced:
-
georanker_serp.create -
georanker_serp.get -
georanker_serp.delete -
georanker_serp.batch_create -
georanker_serp.batch_get -
georanker_serp.compare_locations -
georanker_keywords.create -
georanker_keywords.get -
georanker_keywords.delete -
georanker_keywords.batch_create -
georanker_keywords.batch_get -
georanker_keywords.suggest -
georanker_domain.whois -
georanker_domain.technologies -
georanker_regions.list -
georanker_account.get -
georanker_health.check
Legacy tool names
Set GR_ENABLE_LEGACY_TOOL_NAMES=true or run with --legacy to also expose the old snake_case tools.
Output Format
All tools return compact JSON shaped like:
{
"ok": true,
"action": "georanker_serp.get",
"generated_at": "2026-02-19T12:00:00.000Z",
"request": { "id": "..." },
"data": { "...compact preview..." },
"links": { "raw_resource": "georanker://serp/..." }
}
If a tool fails, it returns ok:false and the MCP tool result will include isError:true.
Publishing
npm
- Build:
npm run build - Login:
npm login - Publish:
npm publish --access public
Smithery MCP Registry
URL publish (recommended for hosted servers):
- Make sure Streamable HTTP is available at
/mcp - Ensure server card exists at
/.well-known/mcp/server-card.json
Then publish from the Smithery UI or via CLI:
smithery mcp publish "https://your-host/mcp" -n @lucas111112/georanker-mcp
Local publish (stdio):
smithery mcp publish --name @lucas111112/georanker-mcp --transport stdio
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。