Best MCP Servers for SEO in 2026

The best MCP servers for SEO in 2026 are the official servers from the SEO tools you already pay for: Ahrefs, Semrush, SE Ranking, Serpstat, DataForSEO and Similarweb for search data, and Screaming Frog for crawls. Google's own Analytics server covers traffic. For Google Search Console, the best-known option is still community-built.

MCP (Model Context Protocol) is the open standard that lets an AI app such as Claude or ChatGPT call outside tools. Connected to an SEO tool, it lets you ask "Which keywords did we lose in the top 10 this month?" or "Crawl this site and list pages with missing titles" without exporting a spreadsheet first.

We checked every entry against the vendor's own docs, GitHub page or pricing page on 2 November 2026. We did not test these products. For a wider marketing stack (ads, CRM, email, CMS), see our list of the best MCP servers for marketing.

How we chose

We used the MCP Guide's standard listicle criteria:

• Live and public. A working MCP server documented on the vendor's site or official GitHub repository, opened on 2 November 2026.
• Clear pricing. Free and open source, included in a published plan, or a clear way to get pricing. Prices come only from official pricing pages.
• Security basics. The docs explain how you sign in (OAuth, API key or service account) and what the server can reach.

We preferred servers run by the tool's maker and labeled the one community project clearly. We dropped Moz and Bing Webmaster Tools because we found only third-party wrappers, not servers from Moz or Microsoft. The order below is not a ranking. Browse more lists on the MCP Guide's best-of hub.

SEO MCP servers compared

1. Ahrefs MCP: backlink and keyword data in your AI app

What it is. Ahrefs says it "provides a hosted MCP server that allows users with a Lite plan or higher to connect their Ahrefs account to supported AI tools." The Ahrefs MCP docs give one Streamable HTTP URL, https://api.ahrefs.com/mcp/mcp, and setup guides for Claude, ChatGPT, Copilot Studio and others.

Best for: SEO teams that already use Ahrefs for backlinks, keyword research and competitor gaps.

Pricing: The Ahrefs pricing page lists "API and MCP access" on Lite ($129 a month), Standard ($249) and Advanced ($449). Enterprise is $1,499 a month with an annual commitment. Each call uses your plan's monthly API units.

Pros:

• Hosted server, so there is nothing to install.
• Keys created through OAuth default to a 200,000-unit monthly cap, which admins can change.
• Connection syncs across devices on the same account.

Cons:

• Starter and Free plans are not listed with MCP access.
• Ahrefs says using the MCP endpoint from "custom scripts, bridges, or standalone HTTP/JSON-RPC clients is unsupported and not permitted."

Checked against: Ahrefs MCP docs on 2 November 2026. Full profile: Ahrefs MCP server.

2. Semrush MCP: SEO, traffic and project data

What it is. Semrush hosts its server at https://mcp.semrush.com/v2/mcp over Streamable HTTP. The Semrush MCP docs say it covers all SEO API endpoints, Trends API reports your plan allows, and read-only methods from the Projects API. Sign-in uses OAuth by default, with an API key fallback.

Best for: Teams that run keyword research, position tracking and competitor traffic checks in Semrush.

Pricing: The Semrush pricing page lists "MCP access" on the Starter plan at $199 a month, or $165.17 a month billed yearly. Pro+ is $299 a month. The docs say eligible plans include 50,000 API units, and extra unit packages are sold separately.

Pros:

• OAuth sign-in with no headers to configure in most clients.
• Long list of documented clients, including Claude, ChatGPT, Cursor and Microsoft Copilot.
• Project write actions are left out, so the AI cannot change your project settings.

Cons:

• Plan names in the docs (Semrush One, SEO Classic) and on the pricing page differ, so confirm your eligibility.
• Every request spends API units, and the cost depends on report type and row count.

Checked against: Semrush MCP docs on 2 November 2026. Full profile: Semrush MCP server.

3. SE Ranking MCP: remote server with four APIs

What it is. SE Ranking says its server "follows the authenticated remote MCP spec, so it's centrally hosted and managed." The SE Ranking MCP page gives the endpoint https://api.seranking.com/mcp and covers its Data, Project, Local Marketing and SE Visible APIs. It notes that "not every API endpoint is available through MCP."

Best for: Small agencies and in-house teams that want rank tracking, local SEO and AI visibility data at a lower price point.

Pricing: The SE Ranking pricing page lists "25K API credits & MCP access" on Core and "100K API credits & MCP access" on Growth. With annual billing, the page shows Core at $103.20 a month and Growth at $223.20 a month (list prices $129 and $279). Enterprise is custom.

Pros:

• OAuth 2.1 with dynamic client registration, or an API key header.
• One-click setup for Claude, Claude Code, ChatGPT, Codex and Cursor.
• Covers local marketing and AI search visibility, not just classic SEO.

Cons:

• API keys carry the same access as your account; restricted or read-only keys "aren't available yet."
• Sub-accounts can't use the one-click sign-in.

Checked against: SE Ranking MCP docs on 2 November 2026.

4. Serpstat MCP: 65 tools from an open-source package

What it is. Serpstat publishes an MCP server on GitHub under its SerpstatGlobal organization, signed by the "Serpstat R&D Team." It installs from npm as @serpstat/serpstat-mcp-server and reads a SERPSTAT_API_TOKEN. The README lists 65 tools across domain, keyword, backlink, URL, project, rank tracking and site audit categories.

Best for: Teams on Serpstat that want rank tracking and site audits inside an AI app.

Pricing: The Serpstat pricing page lists "ChatGPT and Claude Integration (MCP)" on Team ($100 a month) and Team x2 ($169 a month). Agency is $410 a month. The page offers up to 27% off with annual billing. The code itself is MIT licensed.

Pros:

• You can limit tool categories with an environment setting, such as backlinks only.
• Covers audits and rank tracking, not only research data.
• Open-source code you can read before you run it.

Cons:

• Runs locally from npm, so each user installs Node.js 18 or newer.
• Most plans allow one API request per second, and audits consume credits.

Checked against: Serpstat MCP README on 2 November 2026.

5. DataForSEO MCP: raw SEO data, pay as you go

What it is. DataForSEO offers a remote server at https://mcp.dataforseo.com/v3/mcp and a local option through npx -y dataforseo-mcp-server or Docker. Its MCP page says the server connects to the SERP, AI Optimization, Keyword Data, On-Page, Labs, Backlinks, Business Data and Domain Analytics APIs.

Best for: Agencies and developers who want SERP and keyword data at volume without a seat-based SEO subscription.

Pricing: The DataForSEO pricing page describes a "pay-as-you-go pricing model with no subscriptions or recurring fees." The "minimum account top-up is $50" and the balance does not expire. New accounts get a $1 trial credit.

Pros:

• No monthly subscription.
• Remote and local setups, including Docker.
• Modules can be switched off to keep the tool list short.

Cons:

• Every request the AI makes costs money, and broad prompts can trigger many calls.
• The remote server uses a Base64 Basic Auth header, which is a shared secret to protect.

Checked against: DataForSEO MCP page on 2 November 2026. Full profile: DataForSEO MCP server.

6. Similarweb MCP: competitor traffic and search share

What it is. Similarweb's MCP docs describe a server at https://mcp.similarweb.com that brings web, search and app metrics into AI tools. Data includes traffic sources, popular pages, PPC spend, SERP players and SERP clicks.

Best for: Marketers who benchmark competitor traffic and search share before planning content.

Pricing: Pricing on request. The docs say you need "API-only, Business, or Enterprise plans" and an active API key, and that "MCP uses the same data credits as standard API calls."

Pros:

• Competitor data you cannot get from your own analytics.
• Search metrics alongside traffic and audience data.
• Works with Claude Desktop, Cursor, Microsoft Copilot, Zapier and n8n.

Cons:

• No public price for API plans.
• Queries that pull several data types use more credits.

Checked against: Similarweb MCP docs on 2 November 2026.

7. Screaming Frog SEO Spider: your crawler as an MCP server

What it is. Screaming Frog's MCP server guide says the SEO Spider can act as an MCP server in STDIO mode or as a Streamable HTTP server. Tools include sf_crawl, sf_load_crawl, sf_generate_report and sf_generate_bulk_export. The guide recommends version 24.

Best for: Technical SEOs who want an AI app to start crawls and summarize the exports.

Pricing: The guide says "This is a paid licence feature only. It will not work in the free version." The SEO Spider pricing page lists a licence at £199 a year. The free version is limited to 500 URLs.

Pros:

• Uses the crawler you already trust, with its reports and bulk exports.
• HTTP mode runs in the normal app window, so you can watch each action.
• Claude Desktop setup ships as a .mcpb extension.

Cons:

• Optional tools can run Node.js scripts, install packages and write files; sf_write_text_file overwrites without warning.
• Large crawl outputs can fill the AI's context window quickly.

Checked against: Screaming Frog MCP guide on 2 November 2026.

8. Google Analytics MCP server: GA4 reports in plain English

What it is. Google's Analytics MCP docs say the server "can't create, modify, or delete your Google Analytics configuration or settings." It runs locally (pipx run analytics-mcp) and offers standard, realtime, funnel and conversion reports. The GitHub README now also lists one write tool, create_property_annotation, which needs an --allow-write flag and the edit scope.

Best for: Connecting organic traffic and conversions to the keyword data in your SEO tool.

Pricing: Free and open source (Apache 2.0). Normal Google Analytics API quotas apply.

Pros:

• Read-only scope by default.
• Funnel and conversion reports, not just sessions.
• Backed by Google's analytics team.

Cons:

• Local install only; no hosted endpoint is described.
• Setup needs Google Cloud credentials, which is a hurdle for non-technical users.

Checked against: Google Analytics MCP docs on 2 November 2026. Full profile: Google Analytics MCP server.

9. mcp-gsc (community-built): Google Search Console data

What it is. mcp-gsc is an MIT-licensed project by an independent developer, not by Google. It offers search analytics, period comparisons, URL inspection (including batches of up to 10 URLs) and sitemap tools. It signs in with OAuth or a service account.

Best for: SEOs who want to ask questions about Search Console clicks, queries and indexing status.

Pricing: Free and open source (MIT).

Pros:

• Widely used, with about 2.1k GitHub stars when we checked.
• Destructive tools such as delete_site are off unless you set GSC_ALLOW_DESTRUCTIVE.
• URL inspection and indexing checks, not just performance data.

Cons:

• Community-maintained, with no vendor support or security review.
• manage_sitemaps can submit or delete sitemaps, so use a read-only service account for reporting.

Checked against: mcp-gsc GitHub README on 2 November 2026.

10. Microsoft Clarity MCP: behavior data for key pages

What it is. Microsoft publishes the Clarity MCP server as @microsoft/clarity-mcp-server on npm. Its tools query dashboard metrics, list session recordings by filters such as URL and device, and search Clarity docs. It uses a Clarity data export API token.

Best for: Checking how visitors behave on the pages you are trying to rank, such as scroll depth and rage clicks.

Pricing: Free and open source (MIT). Clarity itself is free.

Pros:

• Official Microsoft repository.
• Pairs ranking data with real user behavior.
• Simple token-based setup.

Cons:

• The Data Export API allows 10 requests per project per day and covers only the last 1 to 3 days.
• Local install that needs Node.js.

Checked against: Clarity MCP GitHub README on 2 November 2026.

11. Chrome DevTools MCP: page speed and Lighthouse checks

What it is. Google's Chrome team publishes chrome-devtools-mcp, which lets an AI drive a real Chrome browser. The tool reference lists 60 tools, including lighthouse_audit and performance trace tools such as performance_start_trace.

Best for: Technical SEOs and developers checking Core Web Vitals problems and rendering issues page by page.

Pricing: Free and open source (Apache 2.0).

Pros:

• Runs Lighthouse and performance traces on demand.
• Sees the page as rendered, including JavaScript content.
• Maintained by the Chrome DevTools team.

Cons:

• Usage statistics are on by default (turn off with --no-usage-statistics), and trace URLs may go to the CrUX API (--no-performance-crux).
• Lab checks on one page, not a site-wide crawl.

Checked against: Chrome DevTools MCP README on 2 November 2026.

How to build an SEO stack with MCP

Most teams need one data source, one crawler and one analytics source:

A few guardrails apply to all of them:

1. Set a unit cap. Ahrefs lets admins cap units per MCP key. Do the same in other tools where you can, and watch usage in the first week.
2. Start read-only. Leave Search Console sitemap tools, GA4 annotations and Screaming Frog's Node.js tools off until you need them.
3. Check what the AI reads. SEO tools return a lot of rows. Ask narrow questions to save credits and context.
4. Vet community servers. Read the code and scopes before connecting mcp-gsc to client properties. Our MCP server risk checklist walks through the questions.

For a full workflow, see how SEO teams use MCP.

Next step

If you want help choosing SEO tools for AI search, or want your own product listed and found by AI apps, our MCP marketing team can help you plan it.

Frequently asked questions