Official Data & Analytics

ContextBolt SEO

Pull live keyword research, difficulty scores, and ranking data into any MCP client.

Works with: Claude DesktopClaude CodeCursorWindsurfCline

How to install the ContextBolt SEO MCP server

Add this to your Claude Desktop MCP configuration:

{
  "mcpServers": {
    "contextbolt-seo": {
      "url": "https://seo.contextbolt.app/mcp/YOUR_TOKEN"
    }
  }
}

Add this to your Claude Code MCP configuration:

{
  "mcpServers": {
    "contextbolt-seo": {
      "url": "https://seo.contextbolt.app/mcp/YOUR_TOKEN"
    }
  }
}

Add this to your Cursor MCP configuration:

{
  "mcpServers": {
    "contextbolt-seo": {
      "url": "https://seo.contextbolt.app/mcp/YOUR_TOKEN"
    }
  }
}

Add this to your Windsurf MCP configuration:

{
  "mcpServers": {
    "contextbolt-seo": {
      "url": "https://seo.contextbolt.app/mcp/YOUR_TOKEN"
    }
  }
}

Add this to your Cline MCP configuration:

{
  "mcpServers": {
    "contextbolt-seo": {
      "url": "https://seo.contextbolt.app/mcp/YOUR_TOKEN"
    }
  }
}

ContextBolt SEO is a hosted MCP server that puts live SEO research inside the AI client you already use. One MCP URL gives Claude, Cursor, or any compatible client keyword research, difficulty scores, SERP breakdowns, and competitor analysis on demand. You ask the question in plain English and the numbers come back in the conversation. There is no dashboard to learn and no separate data account to wire up.

You probably found this page in our free MCP server directory. ContextBolt SEO is one of the two products we build ourselves. The directory is free and always will be. The SEO server is the paid tool, at $29 a month for 1,000 lookups.

Why this exists

Serious SEO data has always lived behind dashboards that cost $100 or more a month and pull you out of whatever you were doing. You stop writing, open Ahrefs or Semrush, run a lookup, copy a number, and switch back. The data is good. The workflow is the tax.

ContextBolt SEO removes the tab-switch. The agent you are already prompting becomes the interface to the data. Ask “how hard would it be to rank for iced coffee recipe” and the answer comes back inline: search volume, a difficulty score out of 100, and roughly how many referring domains the current top 10 has. The same class of data as the big tools, asked the way you already think, at a fraction of the price.

What your agent can do with it

Your agent gets six research tools and three backlink tools it can call mid-conversation:

  • keyword_research finds related keywords and their volumes from a seed term.
  • keyword_difficulty scores how hard a term is to rank for, with referring-domain and domain-rank context.
  • serp_overview breaks down who currently ranks on page one for a query.
  • domain_overview summarizes a domain’s organic footprint.
  • ranked_keywords lists the terms a domain already ranks for.
  • competitor_keywords surfaces the domains fighting you for the same terms.
  • backlink_overview pulls a domain’s authority, total backlinks, referring domains and spam score.
  • referring_domains lists every site linking to a domain, ranked.
  • backlink_gap finds the domains linking to your competitors but not you, scored for spam.

The six research tools cost one lookup per call. The three backlink tools cost three each, because backlink data is more expensive to pull.

Because they are MCP tools, the agent chains them. “Find ten low-difficulty keywords around home espresso, check which ones my competitor already ranks for, and tell me the gaps” is one prompt, not ten dashboard sessions.

On top of the research, the memory is automatic. Every lookup is mirrored into a seo-findings folder in your project and kept across sessions, so the agent builds up a living record of your research and never pays to look up the same term twice. You do not call a save command. It just happens.

How it fits with the rest of your stack

ContextBolt SEO is the research layer. It tells you what to target before you have any traffic. Two sibling servers in this directory cover the other half, your own performance:

Run all three in the same client and your agent can move from “what should I write” to “did it work” without ever leaving the conversation.

Gotchas

The plan is 1,000 lookups a month. That is generous for one person doing real research, but heavy automated crawling will burn through it. Top-ups are available and never expire, so you are never forced into a higher recurring tier just to cover a busy week.

The data is keyword and SERP research, not your site’s private analytics. For your own numbers, pair it with the Google Analytics and Search Console servers above. They answer different questions and they do not overlap.

Your MCP URL contains a personal token. Treat it like a password. Do not commit it to a public repo or paste it into a shared config. If it leaks, regenerate it from your ContextBolt SEO account.

Example prompts for the ContextBolt SEO MCP server

What's the keyword difficulty for 'project management software'?

Show me which keywords ahrefs.com ranks for.

Find the domains competing with my site for the same keywords.

ContextBolt SEO MCP server: FAQs

Is ContextBolt SEO free?

No. The MCP server directory you found this on is free, but ContextBolt SEO is a paid product at $29 a month for 1,000 lookups. Top-ups never expire. We kept it to one flat price with no tiers and no per-seat charges.

Which clients can connect to it?

Anything that supports remote HTTP MCP servers. Claude Desktop, Claude Code, Cursor, Windsurf, and Cline all work today. You paste one MCP URL into your client config and the research tools appear in the conversation.

What data does it actually return?

Keyword research, keyword difficulty, SERP overviews, domain overviews, ranked keywords, and competitor keyword analysis. The numbers come from a live SEO data provider, so volumes, difficulty, and SERP positions are real, not estimates.

How is this different from the Google Analytics or Search Console MCP servers?

Those report on traffic your site already has. ContextBolt SEO researches the wider market before you rank: what people search for, how hard a term is, and who you are competing with. They are complementary. Many people run the Google Analytics and Search Console servers alongside this one.

Where does the research get saved?

Automatically. Every lookup mirrors into a seo-findings folder inside your project and persists across sessions, so your agent remembers what it already researched instead of paying to look it up twice. You do not have to ask it to save anything.