How to Connect Porter Metrics to Claude with MCP

What is the Porter Metrics MCP?

An MCP is a way to plug tools into your AI chats. Think of it like a USB port for AI: once a tool has an MCP, any AI chat can connect to it and use its data. And because the standard is universal, the same MCP works everywhere — Claude, Claude Code, ChatGPT, Gemini, Cursor, Antigravity, Lovable, and any other AI tool that accepts APIs or MCPs.

The Porter Metrics MCP brings all your marketing data into those AI tools. The same connectors you use today to send data to Google BigQuery, Google Sheets, or Looker Studio — Google Ads, Meta Ads, TikTok Ads, LinkedIn Ads, Shopify, HubSpot, GA4, Search Console, and 20+ more — are now available inside Claude (and any other MCP-compatible AI).

Why it matters for marketers

With the Porter MCP, you stop waiting on data pulls and start acting on them. Concrete things you can do today:

• Ask questions and get instant answers: “Which Meta Ads campaigns had the highest ROAS last week?” or “Compare Google Ads cost vs. GA4 revenue this month.” Claude pulls the numbers and explains them.
• Spin up dashboards inside the chat: ask for a weekly performance view and get charts, tables, and a summary ready to share — no Looker Studio setup required.
• Generate client-ready reports on demand: “Draft this month’s report for Client X with Google Ads spend, GA4 sessions, and Shopify revenue.” Paste it into Slack, email, or a doc.
• Diagnose performance dips: “Why did our ROAS drop on Tuesday?” Claude looks at your campaigns, channels, and landing pages and explains what changed.
• Build your own marketing apps with vibe-coding tools: combine the Porter MCP with Claude Code, Cursor, Lovable, or Antigravity to create internal tools — lead scoring flows, automated alerts, attribution models — without being a developer.
• Automate recurring work: pair it with scheduled tasks to receive a Monday-morning KPI digest, anomaly alerts, or client updates — unattended.

Prerequisites: what you need before you start

Before you connect Porter to Claude, make sure you have the following ready:

• A Porter Metrics account with your marketing data connected (start a free trial if you don’t have one).
• Claude Desktop or claude.ai (web) — signed in with a plan that supports tool use (all paid plans and the free tier on claude.ai).
• ~5 minutes to complete the full setup.

Step 0 — Connect your marketing data to Porter first

Claude reads your data through Porter, so the first thing you need are your marketing accounts connected inside Porter. If you already use Porter with Looker Studio, Google Sheets, or BigQuery, you can skip this step — every account you connected before is automatically available to Claude too. Connect once, use everywhere.

If this is your first time:

1. Sign in to Porter (or start a free trial).
2. In the top-left corner of the app, click the + Create button.
3. Pick a destination for your data. Porter currently supports four: Looker Studio, Google Sheets, BigQuery, and — new — Claude. It doesn’t matter which one you choose first: any marketing account you authorize here becomes available across all destinations, including Claude via the MCP.
4. Choose the platform you want to connect (Google Ads, Meta Ads, TikTok Ads, GA4, Shopify, HubSpot, etc.) and sign in with the account that owns the data — for example, the Facebook profile that has access to your Meta Ads account, or the Google account that owns your Google Ads.

Once at least one source is connected in Porter, those accounts are ready to use in Claude.

How to connect Porter Metrics to Claude Desktop or Claude Web

This is the easiest path for most marketers. Follow these 10 steps to add Porter as a custom connector in Claude Desktop or claude.ai. The flow is identical on both.

1

Open Claude and click the + inside the chat

Launch Claude Desktop or open claude.ai in your browser. Inside the chat input at the bottom, click the + icon on the left to open the tools menu.

2

Go to Connectors → Manage connectors

In the menu that appears, hover over Connectors. A submenu opens on the right — click Manage connectors to jump into your connector settings.

3

Click the + at the top of the Connectors panel

You’re now in the Customize → Connectors screen. Click the small + icon at the top right of the connector list to start adding a new one.

4

Select Add custom connector

A tiny dropdown appears with two options. Click Add custom connector (the second option).

5

Name your connector Porter

A dialog opens titled Add custom connector (BETA). In the first field (Name), type Porter. This is how you’ll find it inside any conversation later.

6

Paste the Porter MCP URL

In the second field — Remote MCP server URL — paste the Porter endpoint:
https://mcp.portermetrics.com/mcp
You can copy it from the button at the top of this page. Leave Advanced settings untouched.

7

Click Add and sign in with Google

Click the Add button in the bottom right of the dialog. Claude opens your browser on Porter’s sign-in page — click Continue with Google and pick the Google account linked to your Porter workspace (no password needed). Then approve the connection so Claude can read your data.

8

Review Porter’s read-only tools and go back to your chats

After sign-in, Claude shows Porter’s tools: get_blend_id, list_accounts, list_categories, list_data_sources, list_fields, and query_data. These are all read-only — Claude can’t change anything in your accounts. Click back to return to your chats.

9

Start a new chat and prompt anything with your marketing data

In a fresh Claude conversation, type whatever you need. A good first prompt is: “Can you bring my GA4 website sessions by day during this month blended with Search Console clicks?”

10

Get the answer — chart, numbers, and analysis

Claude calls Porter, pulls GA4 sessions and Search Console clicks, blends them, and returns a chart plus a short analysis you can share with your team or client. You’re now chatting with your marketing data.

How to connect Porter Metrics to Claude Code (CLI)

If you use Claude Code in the terminal, the install is a one-liner. This is the recommended path for developers, technical marketers, and anyone automating with scheduled tasks.

Step 1 — Register the Porter MCP

Open your terminal and run the following command:

That registers the Porter MCP in your Claude Code configuration.

Step 2 — Authenticate with Porter

The next time you start a Claude Code session, Claude will open your browser and ask you to sign in to Porter. Click Continue with Google (recommended — no password needed) and pick the Google account linked to your Porter workspace. Complete the sign-in once and the connection persists across sessions.

Step 3 — Verify the connection

Inside Claude Code, run:

You should see portermetrics in the list with status connected. You’re ready to query your data.

How to verify the connection is working

Regardless of which client you used, the fastest way to confirm Porter is connected is to ask Claude a simple discovery question:

Claude should return a list of your accounts (e.g. Google Ads Account 1, Meta Ads Agency Account, GA4 Property). If you see the list, you’re done — the MCP is live and ready for real queries.

What you can do with Claude + Porter Metrics

Once connected, the Porter MCP unlocks these everyday marketing workflows:

Example prompts to try

Copy any of these prompts into a new Claude conversation (with the Porter connector enabled) to see what the integration can do:

Supported data sources and integrations

The Porter MCP covers the same 25+ data sources Porter supports in Looker Studio, Google Sheets, Power BI, and BigQuery:

Considerations, security, and limitations

Before you roll this out to your team or clients, here are the things worth knowing:

Authentication and data access

The MCP uses OAuth 2.0 to connect your Porter account — the standard “Sign in with Google” flow you already use across the web. No passwords are shared with Claude. Every session is scoped to the accounts you authorized in Porter, and Claude can’t see accounts that aren’t linked to your Porter workspace.

Read-only by design

All tools exposed by the Porter MCP are read-only — list_accounts, list_data_sources, list_fields, query_data. Claude cannot modify ad campaigns, pause budgets, or change any settings in your source platforms.

Data privacy

Your data never leaves Porter’s secure pipeline. Claude only receives the specific rows it requests, not your entire account. You can revoke the connector at any time from Claude Settings or from Porter → Integrations.

Rate limits and query size

Queries return up to 1,000 rows per request and respect each platform’s native API rate limits (e.g., Google Ads API quotas, Meta Graph API limits). For large exports or bulk data, use Porter’s Looker Studio, Sheets, or BigQuery connectors instead.

Pricing: how it works

The MCP is already included in your Porter plan — there is no extra fee for connecting Claude. Porter only charges for the data source accounts you connect (for example, one Meta Ads account = one data source account). The same account you already pay for in Looker Studio, Sheets, or BigQuery is the same one you use in Claude — no double billing.

In short: you pay per data source account, and that account works across every destination Porter supports (Looker Studio, Google Sheets, BigQuery, and Claude).

See full pricing details at portermetrics.com/pricing.

Troubleshooting common issues

If something isn’t working, check this list first:

Still stuck? Contact Porter support and include the error message plus the prompt you tried.

Frequently asked questions