seogeoaeo.ai MCP server

seogeoaeo.ai for AI agents

Research competitors, keywords, ads and AI citations, and check any public website for SEO, AEO and GEO readiness, from your coding agent. This repo packages the seogeoaeo.ai MCP server, REST API and agent skill so Claude Code, Cursor, Codex, VS Code and any other MCP client can install them in one step.

Each tool does one job. A site check returns findings with a severity, why each one matters and how to fix it, plus a prioritized list of changes your agent can apply. Re-run the same tool afterwards to verify the fix. A research tool returns the data: competitors, keywords, ads or the pages AI answers cite.

• MCP endpoint: https://seogeoaeo.ai/api/mcp (stateless Streamable HTTP, POST only)
• REST API: https://seogeoaeo.ai/api/v1, described by /openapi.json
• Docs for agents: /agents, /llms.txt, /llms-full.txt

1. Get an API key

Sign in at seogeoaeo.ai, then create a key at seogeoaeo.ai/api-keys. Keys start with sga_. Runs spend credits from your workspace plan, so a plan must be active.

Put the key in an environment variable instead of pasting it into config files:

export SEOGEOAEO_API_KEY=sga_your_key_here

2. Install

Claude Code

claude mcp add --transport http --scope user seogeoaeo https://seogeoaeo.ai/api/mcp \
  --header "Authorization: Bearer $SEOGEOAEO_API_KEY"

Or install the plugin, which adds the MCP server and the skill together:

claude plugin marketplace add SRX9/seogeoaeo-mcp
claude plugin install seogeoaeo@seogeoaeo

The plugin reads SEOGEOAEOAPIKEY from your environment.

Cursor

Add this to ~/.cursor/mcp.json (or .cursor/mcp.json in a project):

{
  "mcpServers": {
    "seogeoaeo": {
      "url": "https://seogeoaeo.ai/api/mcp",
      "headers": { "Authorization": "Bearer ${env:SEOGEOAEO_API_KEY}" }
    }
  }
}

The repo also ships a Cursor plugin manifest in .cursor-plugin/.

VS Code

Add this to .vscode/mcp.json:

{
  "servers": {
    "seogeoaeo": {
      "type": "http",
      "url": "https://seogeoaeo.ai/api/mcp",
      "headers": { "Authorization": "Bearer ${env:SEOGEOAEO_API_KEY}" }
    }
  }
}

Codex

Add this to ~/.codex/config.toml:

[mcp_servers.seogeoaeo]
url = "https://seogeoaeo.ai/api/mcp"
bearer_token_env_var = "SEOGEOAEO_API_KEY"

Claude Desktop and other stdio-only clients

Bridge the remote server with mcp-remote:

{
  "mcpServers": {
    "seogeoaeo": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://seogeoaeo.ai/api/mcp",
        "--header",
        "Authorization:${AUTH_HEADER}"
      ],
      "env": { "AUTH_HEADER": "Bearer sga_your_key_here" }
    }
  }
}

Skill only

If you only want the agent skill, copy skills/seogeoaeo into your agent's skills directory. It teaches the agent to call the REST API with curl, check the credit balance first, retry safely, pick the right research or check tool and apply the returned fixes. The always-current copy is served at seogeoaeo.ai/skills/seogeoaeo/SKILL.md.

What the agent can do

51 tools in two groups, each its own MCP tool named by its slug.

Research answers questions about the market: who your competitors are, what they rank for and which keywords you miss, keyword search volume, difficulty and cost per click, the Google, Facebook, Instagram and LinkedIn ads competitors keep running, and which pages Google AI Overviews and ChatGPT cite for a topic.

Shortened. The full README is on GitHub.