This website uses cookies

Read our Privacy policy and Terms of use for more information.

1. Perplexity MCP

What it does: Real-time sourced research inside Claude — no hallucinations, cited answers. Setup difficulty: Medium (takes ~20 minutes) Docs: docs.perplexity.ai/guides/mcp-server

Step-by-step setup:

  1. Go to perplexity.ai and sign in (or create a free account).

  2. Navigate to Settings > API and generate a new API key. Copy it.

  3. On your computer, locate your Claude Desktop config file:

    • Mac: ~/Library/Application Support/Claude/claude_desktop_config.json

    • Windows: %APPDATA%\Claude\claude_desktop_config.json

  4. Open the file in any text editor (TextEdit on Mac, Notepad on Windows).

  5. Add the following block inside the mcpServers object (create the object if the file is empty):

json

   {
     "mcpServers": {
       "perplexity": {
         "command": "npx",
         "args": ["-y", "@perplexity-ai/mcp-server"],
         "env": {
           "PERPLEXITY_API_KEY": "YOUR_API_KEY_HERE"
         }
       }
     }
   }

  1. Replace YOUR_API_KEY_HERE with the key you copied in step 2.

  2. Save the file, then fully quit and reopen Claude Desktop.

  3. You should now see Perplexity appear as an available tool in Claude.

Note: Node.js must be installed on your computer for npx to work. Download it free at nodejs.org if you don't have it.

Power prompts:

  • "Research the latest developments in [topic] from the last 30 days and give me 5 sourced insights I can use in a LinkedIn post."

  • "Find the top 3 competing perspectives on [trend] with sources and summarise each in 2 sentences."

  • "What are marketers saying about [tool/strategy] right now? Pull me 5 recent opinions or data points with sources."

2. GA4 MCP

What it does: Pulls live Google Analytics data into your Claude conversation without opening a single dashboard. Setup difficulty: Medium (takes ~30 minutes) Docs: Search "GA4 MCP server" on GitHub — several solid community builds available.

Step-by-step setup:

  1. Go to console.cloud.google.com and create a new project (or use an existing one).

  2. Enable the Google Analytics Data API for your project under APIs and Services.

  3. Create OAuth 2.0 credentials: go to APIs and Services > Credentials > Create Credentials > OAuth Client ID. Select "Desktop app" as the application type.

  4. Download the credentials JSON file and save it somewhere safe on your computer.

  5. Find your GA4 Property ID: open analytics.google.com, go to Admin > Property Settings. The ID is the number shown under "Property ID" (e.g. 123456789).

  6. Install a GA4 MCP server via npm in your terminal: npm install -g @modelcontextprotocol/server-google-analytics (or search GitHub for the most current community build).

  7. Add the server to your claude_desktop_config.json file with your credentials path and property ID as environment variables.

  8. Restart Claude Desktop. On first use, Claude will prompt you to complete a Google OAuth login in your browser.

Power prompts:

  • "What were my top 5 traffic sources last week and how do they compare to the previous week?"

  • "Which pages had the highest bounce rate this month — and what might be causing it?"

  • "Show me my conversion funnel drop-off points for the last 30 days and suggest one fix for each stage."

3. Pipeboard MCP

What it does: Connects Meta, Google, TikTok and Snap Ads to Claude in one place — cross-platform campaign data without dashboard hopping. Setup difficulty: Easy (takes ~5 minutes, no technical steps) Docs: pipeboard.co/guides/claude-code

Step-by-step setup:

  1. Go to pipeboard.co and create a free account.

  2. In your Pipeboard dashboard, connect your ad accounts: click "Connect" next to each platform (Meta, Google Ads, TikTok, Snap) and complete the OAuth login for each.

  3. Once connected, go to the MCP section in Pipeboard and copy your unique Remote MCP endpoint URL.

  4. Open Claude Desktop, go to Settings > Developer (or Integrations, depending on your version).

  5. Paste your Pipeboard Remote MCP endpoint into the MCP server URL field. No config file editing required.

  6. Save and restart Claude Desktop. Your ad accounts will now be accessible in Claude.

Power prompts:

  • "Pull my Meta Ads campaigns from the last 7 days. Which ones are burning budget with below-average ROAS?"

  • "Compare my Google Ads and Meta Ads cost-per-lead for the last 30 days. Where should I shift budget?"

  • "Which ad creatives across all my platforms had the highest CTR this month? What do they have in common?"

4. Firecrawl MCP

What it does: Scrapes any live webpage and pulls the content straight into your conversation — no copy-pasting required. Setup difficulty: Technical (takes ~45 minutes, requires config file editing) Docs: docs.firecrawl.dev

Step-by-step setup:

  1. Make sure Node.js is installed on your computer (download at nodejs.org if not).

  2. Go to firecrawl.dev, create an account, and generate an API key from your dashboard.

  3. Locate your Claude Desktop config file:

    • Mac: ~/Library/Application Support/Claude/claude_desktop_config.json

    • Windows: %APPDATA%\Claude\claude_desktop_config.json

  4. Open the file in a text editor and add the following block inside the mcpServers object:

   {
     "mcpServers": {
       "firecrawl": {
         "command": "npx",
         "args": ["-y", "firecrawl-mcp"],
         "env": {
           "FIRECRAWL_API_KEY": "YOUR_API_KEY_HERE"
         }
       }
     }
   }

  1. Replace YOUR_API_KEY_HERE with your Firecrawl API key.

  2. Save the file and fully quit Claude Desktop.

  3. Reopen Claude Desktop. Firecrawl will now appear as an available tool.

Power prompts:

  • "Scrape [competitor URL] and summarise their value proposition, pricing, and key messaging in 5 bullets."

  • "Pull the content from [article URL] and rewrite it as a LinkedIn carousel outline with 8 slides."

  • "Scrape [blog URL] and tell me what topics and keywords they are targeting most consistently."

5. Apify MCP

What it does: Structured data extraction at scale — LinkedIn profiles, Google search results, product listings, all piped into Claude. Setup difficulty: Medium (takes ~20 minutes) Docs: docs.apify.com/platform/integrations/mcp

Step-by-step setup (Remote MCP option — recommended, no config files):

  1. Go to apify.com and create a free account.

  2. In your Apify Console, go to Settings > Integrations > API tokens. Generate a new API token and copy it.

  3. Your Remote MCP endpoint is: https://mcp.apify.com/sse?token=YOUR_API_TOKEN

  4. Open Claude Desktop, go to Settings > Developer (or Integrations).

  5. Paste the endpoint URL (with your token added) into the MCP server URL field.

  6. Save and restart Claude Desktop. Apify actors will now be available in Claude.

Alternative (local config file method):

  1. Complete steps 1-2 above to get your API token.

  2. Add this block to your claude_desktop_config.json:

   "apify": {
     "command": "npx",
     "args": ["-y", "@apify/actors-mcp-server"],
     "env": {
       "APIFY_TOKEN": "YOUR_API_TOKEN_HERE"
     }
   }

  1. Restart Claude Desktop.

Power prompts:

  • "Scrape the top 20 LinkedIn posts about [topic] from the last month and identify the 3 most common content formats used."

  • "Extract all product listings from [e-commerce URL] and give me a pricing comparison table vs my offering."

  • "Pull Google search results for [keyword] and summarise what the top 5 ranking pages are covering that I am not."

6. N8N MCP

What it does: Claude can trigger your N8N automation workflows directly — it does not just analyze, it closes the loop. Setup difficulty: Technical (takes 1+ hour, requires a running N8N instance) Docs: docs.n8n.io

Step-by-step setup (Claude as MCP Client — triggers your N8N workflows):

  1. Open your N8N instance (cloud at n8n.io or self-hosted).

  2. Open the workflow you want Claude to trigger, or create a new one.

  3. Add an "MCP Server Trigger" node at the start of the workflow. Configure it with the tools and descriptions you want Claude to see.

  4. Activate the workflow. N8N will generate an MCP endpoint URL and a Bearer Token — copy both.

  5. Open your claude_desktop_config.json and add:

   "n8n": {
     "command": "npx",
     "args": ["-y", "mcp-remote", "YOUR_N8N_MCP_ENDPOINT_URL"],
     "env": {
       "HEADER_AUTHORIZATION": "Bearer YOUR_BEARER_TOKEN"
     }
   }

  1. Replace the placeholders with your actual endpoint and token.

  2. Save the file, restart Claude Desktop. Claude can now call your N8N workflows by name.

Power prompts:

  • "Trigger my [workflow name] workflow with these inputs: [details] and return the output."

  • "Check the last 5 workflow executions and flag any that failed with a suggested fix."

  • "Run my lead enrichment workflow for [company name] and summarise what you found."

7. Notion MCP

What it does: Makes your entire Notion workspace searchable from Claude — SOPs, briefs, meeting notes, all accessible mid-conversation. Setup difficulty: Easy (takes ~5 minutes, clicks only) Docs: support.claude.com

Step-by-step setup:

  1. Open Claude at claude.ai and sign into your account.

  2. Click on your profile icon in the top right, then go to Settings.

  3. Click on "Integrations" or "Connectors" in the left sidebar.

  4. Find Notion in the list of available connectors and click "Connect."

  5. A Notion authorization window will open. Log into your Notion account if prompted.

  6. Select which Notion pages and databases you want Claude to access, then click "Allow Access."

  7. Return to Claude. Notion is now connected — you can reference your pages in any conversation.

Power prompts:

  • "Search my Notion for everything related to [client or project name] and give me a 5-bullet summary of where things stand."

  • "Find my most recent content strategy doc and pull out the key pillars and next 3 action items."

  • "Look through my Notion meeting notes from the last 2 weeks and create a consolidated list of unresolved action items."

8. Google Drive MCP

What it does: Claude reads your Drive files on demand — no manual uploads, no copy-pasting into the chat. Setup difficulty: Easy (takes ~5 minutes, clicks only) Docs: support.claude.com

Step-by-step setup:

  1. Open Claude at claude.ai and sign into your account.

  2. Click on your profile icon in the top right, then go to Settings.

  3. Click on "Integrations" or "Connectors" in the left sidebar.

  4. Find Google Drive in the list and click "Connect."

  5. A Google authorization window will open. Select the Google account you want to connect.

  6. Review the permissions Claude is requesting (read access to your Drive files) and click "Allow."

  7. Return to Claude. Google Drive is now live — Claude can read any file you reference in conversation.

Power prompts:

  • "Find all documents in my Drive that mention [project or client name] and summarise the key decisions made across them."

  • "Look at my most recent [report type] and tell me what is missing compared to the version before it."

  • "Search my Drive for anything about [topic] created in the last 30 days and give me a quick overview of the key points."

Quick Setup Priority Order

If you are starting from scratch, go in this order:

  1. Google Drive — Easy, immediate value, 5 minutes

  2. Notion — Easy, immediate value, 5 minutes

  3. Pipeboard — Easy via Remote MCP, huge value if you run paid ads

  4. Perplexity — Medium, worth the 20 minutes to configure

  5. Apify — Medium, powerful for research and competitive intel

  6. GA4 — Medium, essential if you manage analytics

  7. Firecrawl — Technical, save for when you are comfortable with config files

  8. N8N — Technical, only if you already have N8N workflows running

Want more Claude breakdowns like this? Follow @rananjayraj on LinkedIn for weekly AI tool guides, marketing automation workflows, and practical Claude tutorials.

LinkedIn | The AI Driven Marketer | Website | Github | Youtube

Keep Reading

© 2026 The AI-Driven Marketer - Proven frameworks, workflows, and strategies for modern marketers..
beehiivPowered by beehiiv