SuperAPI AI Directory

mcp-geo

DigestSEO GEO tracker: measure brand citations and visibility across ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews.

Open website

About this tool

mcp-geo is an AI visibility MCP server launched by DigestSEO, aimed at SEO and GEO (generative engine optimization) scenarios. Its goal is to help users measure brand citations and visibility in AI search and AI chat products, covering seven AI search entry points including ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews.

In terms of main capabilities, mcp-geo performs engine scans based on the configured engine keys, thereby tracking brand citations and visibility. The repository describes it as running locally via stdio, using the user's own API keys, with all data staying on the user's machine. It provides twelve MCP tools; the specific tool names and functional details are not fully expanded in the existing materials.

In terms of usage, Node.js 22.13+ is required (CI covers Node 22 and 24). Users need to set at least one engine key, with optional ones including OPENAI_API_KEY, ANTHROPIC_API_KEY, GEMINI_API_KEY, PERPLEXITY_API_KEY, XAI_API_KEY, SERPAPI_API_KEY; engines without configured keys are safely skipped. It can be run in Claude Desktop or any MCP client via npx: npx -y @digestseo/mcp-geo. Zero keys can also be used for tool discovery, and keys can be added as needed when performing engine scans.

For ChatGPT, which does not support local STDIO MCP servers, a self-hosted remote MCP solution or the OpenAI Secure MCP Tunnel solution is required; Perplexity Computer supports adding custom remote MCP connectors on eligible plans; Replit Agent supports remote HTTPS MCP servers, added via Integrations -> MCP Servers and using an OAuth flow. In addition, the documentation also provides Amp, OpenCode v2, Mistral Vibe Code, LibreChat, AnythingLLM, Langflow, Flowise, Kilo Code, Google Antigravity, Cherry Studio, Raycast AI, Msty Studio, Jan Desktop / Jan Agent, Zed, TraeCode, Qwen Code, Augment Code / Auggie, GitHub Copilot CLI, and other clients and platforms.

Applicable scenarios include: SEO and GEO practitioners who want to understand brand citations and exposure across multiple AI search interfaces; developers integrating brand visibility scanning tools into various MCP clients; teams self-hosting to keep data localized.

In terms of limitations, ChatGPT cannot directly connect to local STDIO MCP servers; the public geo-mcp.digestseo.com/mcp endpoint is not a scanning service that can be used directly without a key; Replit only supports remote HTTPS MCP servers and requires a self-built Worker URL; Copilot cloud agents and code review currently do not support remote MCP servers that rely on OAuth, and code review only accepts tools marked as read-only. Some advanced features and billing details are not yet provided in the existing materials.