SEO Researcher

SEO Researcher gives Claude, Codex, Cursor, and other agents the live search research they are missing before they write: Google competitors, topics, entities, search questions, AI citations, content gaps, and source URLs.

The public website, REST façade, and remote MCP server run on Cloudflare. Research jobs use the existing Rankability research infrastructure behind a narrow, versioned adapter. This repository contains no SERP-provider runtime and no Rankability secrets.

Connect an agent

Use the Streamable HTTP MCP endpoint:

https://mcp.seoresearcher.ai/mcp

The first connection opens OAuth. After authorization, ask your agent:

Research “best CRM software” before we create this article.

The focused tool surface is:

• seo_research_start

• seo_research_status

• seo_research_get

• seo_research_list

Related MCP server: Cold Email & Outreach MCP Server

REST API

The REST façade is available at https://api.seoresearcher.ai/v1. Send the same OAuth access token or Rankability API key as a bearer token. See docs/API.md and openapi.yaml.

Local development

npm install
npm run dev
npm run dev:worker

Copy .env.example to .dev.vars for local Worker secrets. Never commit .dev.vars.

Validation and deployment

npm run validate
npm run deploy

Cloudflare serves the same Worker at seoresearcher.ai, api.seoresearcher.ai, and mcp.seoresearcher.ai. The Worker separates each host's surface, records edge analytics, and proxies only the dedicated SEO Researcher contract.

Security

See SECURITY.md. OAuth is preferred. Raw access tokens are never written to analytics or logs.

License

The public integration and deployment shell is available under the MIT License. Rankability's private research engine is not part of this repository.