SEO API: a REST API and an MCP server for your sites' search data

XeoRank's SEO API is a REST API at https://api.xeorank.com that returns your sites' Search Console queries, competitor pages, content gaps, page analysis and crawl data as JSON, authenticated with a personal API key from your settings. For AI assistants, the same platform is exposed as an MCP server at https://mcp.xeorank.com/mcp with more than 200 tools, and the full reference for both lives in the docs.

The two serve different callers. Your code calls the REST API. Your AI assistant calls the MCP server.

Who it is for

• Developers who want their sites' Search Console data and SEO checks inside an internal dashboard, a data pipeline or a script.
• Teams that want an SEO check in CI, for example crawling a staging URL and failing the build when the title or meta description is missing.
• Agencies pulling client data into their own reporting.
• Anyone building with Claude, ChatGPT, Codex or Cursor who wants the assistant to run SEO work directly, through the MCP server.

Authentication

Create a personal API key in your XeoRank settings. Each key shows its prefix, its creation date and when it was last used, and you can revoke it whenever you like. Send it with every request in the X-Agent-Key header. The CLI uses the key saved by xyle login; if you have not logged in, it falls back to a key in the AGENT_API_KEY environment variable.

A personal key is bound to you, and a request that names a different account is refused. Treat it like a password: keep it in an environment variable or your CI secret store rather than in source control, and revoke it from settings if it leaks. The XeoRank CLI calls this same API, so a key that works in one works in the other.

A request, start to finish

The API reference documents the request shape. Listing a site's top Search Console queries looks like this:

GET https://api.xeorank.com/queries?site=your-domain.com&limit=5&email=you@example.com
X-Agent-Key: your-api-key

The response is a JSON array of objects with query, impressions, clicks, ctr and position.

Documented endpoints

These are the endpoints in the public reference at /docs:

The /rewrite endpoint returns suggestions. It does not change your page.

For background on what the gap and rewrite data mean, see content gap analysis and content briefs.

What people build with it

• A query dashboard. Call /queries from your own reporting job and chart clicks and position next to your product metrics.
• A pre-publish check. Call /crawl on a freshly deployed public URL and fail the build if the title, meta description or headings are missing.
• A refresh list. Call /gaps for your key pages and send the ones with missing topics to your content team.
• A Search Console health check. Call /sites/{id}/search-console to read coverage, sitemaps and indexing issues, and alert someone when errors appear.

Each of these runs when your code calls it, on whatever trigger you choose.

The MCP server: the rest of the platform

The REST reference covers a core set of data. Much of XeoRank's work, including full-site crawls, keyword research and clustering, rank tracking, AI visibility checks, GA4, content briefs and WordPress editing, is exposed through the MCP server.

Connect it to your assistant with the URL https://mcp.xeorank.com/mcp:

• Claude, Cursor and VS Code sign in with OAuth, so there is no key to copy.
• ChatGPT on the web connects it as a custom connector with OAuth sign-in, after you turn on developer mode under Settings and then Security.
• Codex can use OAuth like ChatGPT, or a bearer API key: bearer_token_env_var in ~/.codex/config.toml names an environment variable that holds your key.

The server is listed in the official MCP Registry as com.xeorank/xeorank. The MCP setup guide has the exact steps for each client, and the SEO MCP server page covers what you can ask it to do.

What the API does not do

• The documented REST endpoints do not edit your website. Site edits go through the MCP server. WordPress writes are off for each site until you enable them.
• It is not a bulk keyword or search results data feed. It returns data for the sites in your account.
• The documented REST endpoints do not cover everything. AI visibility, rank tracking, keyword research and briefs are reached through the MCP server, not through the public REST reference.
• There is no separate client library to install. Any HTTP client works, and the npm CLI is a ready-made client for the terminal.
• Requests are rate limited, so batch work should pace itself rather than fire requests in parallel.

Pricing

API and MCP access come with every plan. Starter is $19 a month for one site, with a 14-day free trial and no card to start. Pro is $49 a month and Growth is $109 a month. See pricing for what each plan includes.