MCP (Model Context Protocol) connectors let Claude pull live data from your SEO tools, CRM, analytics platforms, email systems, and ad accounts inside a single conversation. No CSV exports. No tab switching. No copy-pasting data into prompts. This is the complete independent guide to every marketing MCP connector available today, how to set each one up, and the real workflows that replace hours of manual reporting.

The average marketing team uses 12 different tools. Every cross-platform analysis starts the same way: export data from tool A, open tool B, export again, paste both into a spreadsheet, and then manually summarize the findings before you can even start making decisions.

MCP eliminates that entire workflow. With MCP connectors, Claude connects directly to your marketing tools and pulls live data in real time. You ask Claude "show me which blog posts lost the most traffic in the last 90 days, what keywords they rank for in Ahrefs, and how many leads those posts generated in HubSpot" and you get one consolidated answer from one conversation (Ahrefs, What Is an MCP Server).

MCP stands for Model Context Protocol. Anthropic created it as an open standard that lets AI assistants connect securely to external tools and data sources. Instead of AI tools guessing or hallucinating data, MCP gives them a direct, authenticated connection to live information from the tools you already pay for (Anthropic, Claude Code MCP Documentation).

Our guide covers every marketing-relevant MCP connector available right now, grouped by function. Each section includes the exact setup steps, the specific marketing workflows each connector enables, and the combinations that produce the highest-leverage results. For how to use Claude for marketing more broadly, see our complete Claude for marketing guide. For setting up Claude Projects that remember your brand context, see our Claude Projects setup guide.

What Are MCP Connectors and How Do They Work for Marketing Teams?

An MCP connector is a secure bridge between Claude and an external tool. When you connect Ahrefs via MCP, Claude does not guess at your SEO data. It calls the Ahrefs API, retrieves your actual keyword rankings, backlink profiles, and traffic data, and uses that real information to answer your question.

The architecture has three parts. The MCP host is your AI assistant (Claude). The MCP client is a connector that the host creates for each server. The MCP server is a program that exposes your tool's data and functions to the AI (Ahrefs, What Is an MCP Server). Authentication, rate limits, and error handling are all managed by the protocol.

For marketing teams, this changes three things:

1. Cross-platform analysis happens in one conversation: Instead of exporting from Semrush, GA4, HubSpot, and Search Console separately, Claude queries all four and synthesizes the answer. One question replaces four tool logins and a spreadsheet.

2. Claude works with your actual data, not generic advice: When you ask "which content should I update first?" Claude pulls your real traffic trends, your real keyword positions, and your real conversion data, and ranks your content by actual impact. No generic recommendations.

3. Actions happen inside the conversation: Claude does not just report data. With write-enabled connectors like HubSpot, Claude creates contacts, updates deal stages, logs activities, and assigns tasks directly from the chat. According to HubSpot's documentation, the connector supports creating and updating CRM records, logging activities and notes, and generating insights from engagement history including emails, calls, and meetings (HubSpot, Claude Connector Setup Guide).

For understanding how MCP-connected Claude fits into your broader AI search strategy, see our GEO guide.

Every Marketing MCP Connector Available Right Now (Grouped by Function)

SEO and search analytics connectors

Ahrefs launched an official remote MCP server that connects directly to Claude and ChatGPT. You authenticate with your Ahrefs API key (requires an active Ahrefs subscription with API access) and Claude can pull keyword data, backlink profiles, domain ratings, and competitive analysis in real time (Ahrefs, Official MCP Server). The Ahrefs remote HTTP endpoint (https://api.ahrefs.com/mcp/mcp) is the recommended setup for Claude Code users.

Semrush has both an official connector in Claude's directory (connect via OAuth in Settings > Connectors on Claude.ai) and a community-maintained MCP server on GitHub with 77 tools covering domain analytics, keyword research, backlinks, traffic analysis, and competitive intelligence (GitHub, mrkooblu/semrush-mcp). The community server requires a Semrush API key and runs locally via stdio. The official Semrush guide details the setup process for connecting to Claude, ChatGPT, Cursor, and VS Code (Semrush, MCP Connector Guide).

DataForSEO offers a pay-per-query alternative. Instead of monthly subscriptions (Ahrefs starts at $129/month, Semrush at $139.95/month), DataForSEO charges per API call with a $50 minimum deposit (NextGrowth.ai, DataForSEO MCP Server Setup). The MCP server supports keyword research, SERP analysis, backlink data, and on-page analysis.

For how these SEO tools feed into your content research workflow, see our keyword intent analysis guide. For the entity-based approach to keyword research, see our keyword clustering for entity-based AI search guide.

CRM and sales connectors

HubSpot offers the most mature marketing CRM integration. The connector in Claude's directory uses OAuth authentication. Once connected, Claude can get AI insights from your HubSpot records, create and update contacts and deals, log activities and tasks, and generate insights from engagement history. HubSpot admins control organization access through Settings > Integrations > Connected Apps (HubSpot, Claude Connector Setup).

For teams that need deeper HubSpot integration (custom objects, workflows, automation), HubSpot also offers a Developer MCP server for building custom applications and a remote MCP server for programmatic CRM access via Private App API keys (HubSpot Developers, MCP Server).

Clay is particularly valuable for marketing teams doing competitive research and prospect enrichment. Connect it and ask Claude to "find 5 companies that compete with [client domain], identify who leads their marketing, and what tools they are using." Amplemarket's MCP integration enables similar prospect research and list building workflows (Amplemarket, MCP for Marketing).

For how CRM data connects to your content funnel, see our SEO funnel strategy guide. For using Claude to build lead generation content, see our gated content lead generation guide.

Analytics and data connectors

Windsor.ai is the Swiss Army knife for marketing analytics. It connects Claude to 325+ marketing data sources including Meta Ads, Google Ads, GA4, Instagram, Shopify, LinkedIn Ads, and more through a single MCP endpoint. This means one connector gives Claude access to your entire paid media stack, your web analytics, and your ecommerce data simultaneously (Windsor.ai, Claude Connectors). For marketing teams running multi-channel campaigns, this eliminates the biggest bottleneck: consolidating data from different ad platforms into one view.

Google BigQuery via MCP lets Claude query your data warehouse directly, which is critical for teams that centralize marketing data in BigQuery for cross-platform analysis.

For tracking AI traffic sources in your analytics, see our GA4 guide for AI traffic.

Content, CMS, and design connectors

For how to structure content for both traditional SEO and AI search visibility, see our topic clusters guide. For optimizing content with structured data, see our schema markup guide.

Email marketing connectors

Communication and project management connectors

Automation connectors

For building automated content workflows, see our content marketing strategy guide.

How to Set Up MCP Connectors on Claude.ai (Step by Step)

There are two setup paths depending on whether you use Claude.ai (the web/app interface) or Claude Code (the command line tool).

Setting up connectors on Claude.ai (no code required)

This is the path for most marketing teams. No terminal commands. No API keys (for directory connectors).

Step 1: Open Claude.ai. Click the "Customize" button in the sidebar. Click "Connectors."

Step 2: Click the plus (+) icon. Select "Browse connectors" to see the official directory.

Step 3: Find the connector you want (HubSpot, Slack, Google Drive, etc.). Click "Connect."

Step 4: Follow the OAuth authentication flow. This typically opens a new browser tab where you log in to the external service and grant Claude permission to access your data.

Step 5: Return to Claude. The connector icon appears at the bottom of your chat window, confirming the connection is active.

Every connector in Claude's directory uses OAuth authentication, which means you never share API keys or passwords with Claude directly. You authorize access through the tool's own login system, and permissions can be revoked at any time from the connected tool's settings (XDA Developers, Claude Connectors Guide).

Custom connectors (for tools not in the directory) require a remote MCP server URL. Click "Add custom connector" instead of "Browse connectors," enter the connector name and the MCP server URL, and click Connect (Anthropic, Custom Connectors Help).

Setting up connectors on Claude Code (command line)

For technical marketing teams using Claude Code, the setup is one command per connector:

claude mcp add --transport http [name] [url]

claude mcp add --transport http [name] [url]

claude mcp add --transport http [name] [url]

For example, to add Ahrefs:

To add the community Semrush MCP server (requires API key):

Set the environment variable: export SEMRUSH_API_KEY=your_api_key_here

To verify your connectors are active, run /mcp inside Claude Code to see all connected servers and their status.

For the complete Claude Code reference, see Anthropic's MCP documentation.

5 High-Impact Marketing Workflows Using MCP Connectors

These workflows combine multiple connectors to replace manual processes that typically take hours.

Workflow 1: Automated content decay detection and prioritization

Connectors used: Ahrefs (or Semrush) + Google Search Console + HubSpot

The prompt: "Find all blog posts that dropped more than 30% in traffic over the last 3 months using Google Search Console. For each one, pull the current keyword rankings from Ahrefs. Cross-reference with HubSpot to see how many leads each post generated last quarter. Prioritize by which posts lost the most leads, not just the most traffic."

What this replaces: Exporting GSC data, exporting Ahrefs rankings, exporting HubSpot attribution data, merging in a spreadsheet, and manually sorting. Typically 2-3 hours of work.

For the content refresh strategy these insights feed into, see our content freshness and decay guide.

Workflow 2: Competitive content gap analysis

Connectors used: Ahrefs (or Semrush) + Similarweb

The prompt: "Pull the top 50 keywords that [competitor domain] ranks for in the top 10 that we do not have first-page rankings for. For each keyword, show the search volume, keyword difficulty, and the URL that ranks. Then pull traffic estimates from Similarweb for their top 10 pages to see which ones drive the most visits."

What this replaces: Running content gap reports in Ahrefs, cross-referencing with Similarweb traffic data, and building a prioritized opportunity spreadsheet. Typically 1-2 hours.

For the link building strategies these gaps reveal, see our ecommerce link building guide.

Workflow 3: Multi-channel campaign performance summary

Connectors used: Windsor.ai (connects to GA4, Google Ads, Meta Ads, LinkedIn Ads, etc.)

The prompt: "Summarize KPIs across all channels for the last 30 days. Show spend, conversions, cost per conversion, and ROAS for Google Ads, Meta Ads, and LinkedIn Ads. Identify the top 3 underperforming campaigns and the top 3 overperforming campaigns by ROAS."

What this replaces: Logging into Google Ads, Meta Ads Manager, and LinkedIn Campaign Manager separately, exporting data from each, normalizing metrics, and building a comparison. Typically 1-2 hours weekly.

Workflow 4: Lead source analysis and content attribution

Connectors used: HubSpot + Google Search Console

The prompt: "Count all new contacts in HubSpot from the last 6 months where the source field includes organic search. Pull the landing page URLs for the top 20 organic lead sources. Cross-reference with Search Console to show which keywords drive traffic to those pages and their current ranking positions."

What this replaces: Running HubSpot attribution reports, matching landing pages to GSC keyword data, and building a content-to-leads attribution map. Typically 3-4 hours.

For optimizing the landing pages these leads arrive on, see our landing page optimization guide.

Workflow 5: AI search visibility audit

Connectors used: Ahrefs + any CMS connector (Notion, WordPress, Webflow)

The prompt: "Pull our top 50 organic pages by traffic from Ahrefs. For each, check if the page has FAQ schema, Article schema, and clear question-based H2 headings. Flag pages that rank in positions 4-20 but lack structured data, because these are the highest-leverage pages for AI Overview citations."

What this replaces: Manual audit of schema markup, heading structure, and AI-readiness across your top content. Typically a full day of work.

For the GEO optimization framework this audit supports, see our complete GEO guide. For the source gap analysis methodology, see our source gap analysis for AI search guide.

Claude MCP Connector Pricing: What You Actually Pay

Every built-in connector in Claude's directory is free to use. You only need a paid Claude plan and an active subscription to the connected tool.

The cost of the connected tools is separate. Ahrefs starts at $129/month with API access. Semrush starts at $139.95/month. HubSpot's free CRM works for basic connector use. Windsor.ai starts with a 30-day free trial.

DataForSEO offers a pay-per-query model starting with a $50 minimum deposit, which is worth considering if you need SEO data but do not need the full Ahrefs or Semrush subscription (NextGrowth.ai, DataForSEO MCP Setup).

For help deciding which tools to invest in, see our SEO tools comparison.

MCP Security: What Marketing Teams Need to Know

MCP connectors use OAuth 2.0 authentication for directory connectors, which means Claude never sees your passwords. You authenticate through each tool's own login system. Permissions can be revoked from the connected tool's settings at any time.

Key security considerations for marketing teams:

• Start with read-only access: Most connectors support both read and write actions. Start with read-only to test the integration before enabling write actions like creating contacts or updating deals.

• Review permissions carefully: HubSpot's connector documentation notes that admins control organization access and that giving users access cannot be undone: even if the connector is uninstalled, users retain access until they disconnect individually (HubSpot, Connector Setup).

• Custom connectors are not verified by Anthropic: Only connectors in the official directory have been reviewed for quality and security. Custom connectors using third-party MCP server URLs should only be used from trusted sources (Anthropic, Custom Connectors Help).

• HubSpot respects data center location: If your HubSpot account is hosted in the EU, requests are routed through HubSpot's EU data center. Once data reaches Anthropic's platform, your Anthropic license terms govern data processing.

For understanding how AI tools handle your marketing data more broadly, see our AI ethics in marketing guide.

Set Up Your First Marketing MCP Connector Today

Open Claude.ai. Click Customize. Click Connectors. Connect one tool you use daily. Start with Google Drive if you are not sure where to begin, because it gives Claude access to every document, spreadsheet, and presentation your marketing team has already created, with zero new data to input.

Then try one cross-platform workflow. Connect Ahrefs and ask Claude to identify your content decay opportunities. Connect HubSpot and ask Claude to summarize your pipeline by deal stage. Connect Windsor.ai and ask Claude to compare your ad spend efficiency across channels.

The marketing teams that build their MCP connector stack now are compounding an operational advantage. Every hour they save on manual data consolidation is an hour spent on strategy, creative, and execution. That gap widens every week.

If you need expert help building a Claude-connected marketing system for your team, Passionfruit's AI marketing consultants set up MCP workflows, Claude Projects, and AI-optimized content strategies for SaaS and ecommerce brands. See our case studies for measurable results.

Frequently Asked Questions

What is MCP in the context of Claude and marketing tools?

MCP stands for Model Context Protocol. It is an open standard created by Anthropic that lets AI assistants like Claude connect securely to external tools and data sources. For marketing teams, this means Claude can pull live data from your SEO tools, CRM, analytics platforms, and email systems inside a single conversation, instead of requiring you to export data manually. The protocol handles authentication, rate limits, and data formatting automatically.

How do I connect Claude to HubSpot using MCP?

Open Claude.ai, click Customize in the sidebar, then click Connectors. Browse the connector directory and select HubSpot. Click Connect and complete the OAuth authentication flow by logging into your HubSpot account. Once connected, Claude can access your contacts, deals, engagement history, and pipeline data. HubSpot admins must first enable the connector through Settings > Integrations > Connected Apps > HubSpot connector for Claude and click "Give users access."

Can I connect Claude to Semrush for keyword research?

Yes. Semrush is available both in Claude's official connector directory (connect via OAuth) and as a community-maintained MCP server with 77 tools covering domain analytics, keyword research, backlinks, traffic analysis, and competitive intelligence. The official connector requires only an OAuth login. The community server on GitHub requires a Semrush API key and provides deeper access to Semrush's full API.

Does connecting MCP tools to Claude cost extra?

No. All connectors in Claude's directory are free to use. You need a paid Claude plan (Pro at $20/month minimum) and an active subscription to the tool you are connecting. Claude does not charge per connector or per API call. However, some connected tools may have API usage limits based on your subscription tier.

Is my marketing data safe when using MCP connectors?

Directory connectors use OAuth 2.0 authentication, meaning Claude never stores your passwords. You authorize access through each tool's login system and can revoke permissions at any time. Custom connectors are not verified by Anthropic and should only be added from trusted sources. For enterprise teams, Claude Team and Enterprise plans offer admin controls, audit logs, and the ability to restrict which connectors team members can use.

What is the difference between Claude.ai connectors and Claude Code MCP servers?

Claude.ai connectors use the web interface: you click Connect, authenticate via OAuth, and start using the tool in conversation. Claude Code MCP servers use the command line: you run claude mcp add with a server URL, and the connection is available in your terminal session. Both use the same MCP protocol underneath. Claude.ai connectors are easier for non-technical marketers. Claude Code is more powerful for technical teams building automated workflows.

Which MCP connectors should marketing teams set up first?

Start with the three highest-impact connectors for most marketing teams: Google Drive (Claude searches and reads your documents, spreadsheets, and presentations), Gmail (Claude searches emails and drafts replies), and one data connector for your primary marketing platform (HubSpot for CRM-driven teams, Ahrefs or Semrush for SEO-driven teams, Windsor.ai for paid media teams). Add more connectors as you identify specific cross-platform workflows that benefit from live data access.