The PMA MCP server is currently in Alpha. Access is granted per organization on request. To request access, contact your PMA representative or open a support ticket with the subject line "MCP Early Access" at support.powermyanalytics.com.

What Is the PMA MCP Server?

The Power My Analytics MCP server is a hosted endpoint that connects MCP-compatible AI assistants (Claude Desktop, Claude Code, Claude Web, ChatGPT, and others) directly to your PMA Hub. Once connected, you can ask questions about your marketing data in plain language and get answers grounded in your warehoused PMA data, without first having to build a report.

Think of it as a conversational analytics surface for your PMA Hub. Instead of building a Looker Studio scorecard or assembling a Data Builder dataset to answer a one-off question, you can simply ask.

How the MCP Server Fits in Your PMA Workflow

The MCP server is a new way to consume the data you have already centralized in your PMA Hub. It complements your existing reporting tools rather than replacing them:

• Looker Studio, Google Sheets, and Microsoft Excel reports stay authoritative for recurring, shareable, board-ready dashboards.
• Data Builder remains your tool for saving multi-source, multi-table reporting workspaces. The MCP server can read, write, and query Data Builder datasets through dedicated tools, but it does not replace the Data Builder UI.
• Sources, Schema Explorer, and Manage Accounts in the Hub remain the way you configure connections, sub-accounts, and sync schedules.
• The PMA API and Custom-plan exports continue to serve programmatic, scheduled, and large-volume access patterns.

The MCP server shines when the question is exploratory or ad-hoc, like "How is my Facebook Ads performance compared to last month?" or "Did anything weird happen this week?" or "Which campaigns are driving the most conversions?" You ask, the AI answers, without first having to build a report.

Where the MCP Server Fits in the AI Landscape

The Model Context Protocol (MCP) is an open standard for connecting AI assistants to external data sources and tools. The PMA MCP server is Power My Analytics' implementation of that standard: it makes your warehoused marketing data available to any MCP-compatible AI assistant, so the AI can read it and answer questions about it on demand.

That makes the PMA MCP server complementary to (and different from) two adjacent kinds of AI tooling:

• Single-platform third-party MCPs (for example, a YouTube MCP, a Shopify MCP, or a Funnel/Supermetrics MCP) typically expose live data from one specific platform. The PMA MCP queries the warehoused, blended, multi-platform layer in your Hub, so cross-platform comparison is a single tool call rather than a manual reconciliation.
• Hosted AI analytics tools that ingest your data into their own warehouse and analyze it. The PMA MCP server keeps your data in PMA's existing warehouse; the AI assistant only receives the specific query results it needs to answer your question.

Who Can Use the PMA MCP Server

On the PMA side, you will need:

• An active PMA Hub with at least one connected data source that has data syncing.
• Org Admin role in the PMA Hub.
• Alpha access granted by Power My Analytics. To request access, contact your PMA representative or open a support ticket with the subject line "MCP Early Access" at support.powermyanalytics.com.

On the integration side, the PMA MCP server supports two equivalent paths depending on what you are building:

For AI assistant clients (browser-based OAuth):

• A Claude or ChatGPT plan tier that supports custom MCP connectors (e.g., Claude Pro or Team), or any other MCP-compatible client that supports Streamable HTTP transport and OAuth 2.1.
• Permission within your AI client to add custom MCP connectors. Individual users on a qualifying plan can add the connector themselves; in organization contexts, the org Owner adds it once for the whole org.

For headless / server-to-server integrations (no browser, no interactive user session):

• An MCP client library that supports Streamable HTTP transport with a custom Authorization header (for example, the official @modelcontextprotocol/sdk package, or any equivalent library in your language of choice).
• Your existing PMA API token, which the MCP server accepts directly as a Bearer credential. No OAuth handshake is required for this path. See Headless/Server-to-Server Integration with PMA MCP Server for the full walkthrough.

What You Can Do with It

Once connected, your AI assistant can browse your connected data sources, inspect schemas, run analytics queries against your warehoused marketing data, and read or build Data Builder datasets, all through natural-language conversation. Common use patterns include:

• Cross-platform performance comparisons. "Compare last month's spend across Facebook Ads and Google Ads."
• Anomaly investigation. "My total ad spend dropped this week. Where did it drop, and is it a real drop or a data gap?"
• Quick dataset and data-table builds. "Create a dataset called 'Paid Spend Weekly' that blends Facebook Ads and Google Ads, with spend, impressions, and clicks for the last 30 days."
• Connector and account health checks. "Why isn't my data syncing?"
• Schema exploration. "What fields can I query for HubSpot?"

The PMA MCP server exposes 22 first-party tools across five families (discovery and diagnostics, sub-account and activity, analytics, dataset/Data Builder, and template browsing), plus one slim MCP protocol resource for client-side caching.

Advantages over Other Reporting and Analytics Tools

• No new credentials. Browser-based AI clients authenticate against your existing PMA login (Google SSO or email + password) via OAuth 2.1 with PKCE. Headless integrations use your existing PMA API token directly as a Bearer credential. Either path works without standing up a separate identity for the MCP server.
• No data migration. Your data stays in PMA's existing data warehouse; the AI client receives only the specific query results it needs.
• Cross-platform out of the box. Because the PMA MCP queries the warehoused, deduplicated, time-zone-normalized SQL layer, cross-platform comparisons are immediate, not manual.
• Editable across surfaces. Datasets created via MCP are the same Data Builder datasets visible in the Hub UI; you can edit either one in the other surface.
• Conversational instead of click-based. Most PMA workflows answer recurring reporting needs. The MCP server is designed for the ad-hoc questions in between.

How to Begin Using the PMA MCP Server

1. Request Alpha access. Open a support ticket with the subject line "MCP Early Access" at support.powermyanalytics.com, or contact your PMA representative.
2. Wait for confirmation. Once your organization is allow-listed, PMA support will follow up with confirmation that your org is provisioned and the server URL.
3. Connect your client. For an AI assistant, follow the walkthrough for your client of choice (linked below). For a headless / server-to-server integration, follow Headless/Server-to-Server Integration with PMA MCP Server instead.
4. Ask a data question (or call a tool). Once connected, just ask. The AI client picks the right tool automatically based on your prompt; in a headless integration, your service calls the tools directly through your MCP client library.

Learn More

• Power My Analytics MCP Server Guide: the full overview of the server, its tools, and worked examples covering both basic and advanced use.
• PMA MCP Tools Reference: every tool exposed by the server, with descriptions, typical prompts, and key behaviors.
• Connect to the PMA MCP Server from Claude Desktop
• Connect to the PMA MCP Server from Claude Code
• Connect to the PMA MCP Server from ChatGPT
• Headless/Server-to-Server Integration with PMA MCP Server: connect a custom backend service (Node.js, Fly.io, scheduled jobs, agentic systems) to the PMA MCP server using a static Bearer token, no browser required.
• Troubleshooting and Important Considerations for PMA MCP Server: known limitations, current alpha behaviors, and resolution steps for common issues.

Have feedback on the PMA MCP server, an idea for a new tool, or a use case you would like us to write up? We would love to hear it. Reply to your early-access ticket or contact contactus@powermyanalytics.com with the subject line "MCP Feedback".