MCP - Mixpanel Docs (original) (raw)

Mixpanel MCP Integration

Overview

The Mixpanel Model Context Protocol (MCP) integration enables organizations to analyze product and replay data using conversational AI. Teams can query events, session replays, funnels, and user behavior through AI interfaces (Claude, Cursor, etc.) using natural language, eliminating the need for complex dashboard navigation or technical query languages.

Mixpanel MCP Tools

The MCP integration provides access to all core Mixpanel analysis capabilities through natural language queries:

Analytics Functions

Segmentation: Event counts and unique users with filtering and grouping (run_segmentation_query)
Funnels: Conversion rate analysis across user journeys (run_funnels_query)
Retention: User engagement and return behavior tracking (run_retention_query)
Frequency: User engagement patterns and usage intensity (run_frequency_query)

Data Discovery & Exploration

Project Management: Access projects and workspace information (, )
Event Discovery: Browse all available events in your project ()
Event Properties: Explore properties available for each event ()
Property Values: Discover specific values for event properties ()
Data Quality: Monitor and detect volume anomalies in data ingestion ()
Lexicon Integration: Get direct links to event definitions and documentation in Mixpanel Lexicon ()

Qualitative Analysis

Session Replays: Analyze a specific user’s replays in combination with event data ()

Query Capabilities

Natural Language Filtering: Use complex boolean expressions for data filtering
Advanced Grouping: Segment data by properties with mathematical operations
Time-based Analysis: Analyze trends across different time periods (hourly, daily, weekly, monthly)
Numerical Aggregations: Perform sum, average, and bucket analysis on numeric properties

For more information on organizing and documenting your events, see Mixpanel Lexicon documentation.

Getting Started

Beta Access

The Mixpanel MCP integration is currently in Beta. You can start using it right away by following the Implementation Guide below.

Permissions & Access Control

Admin Configuration Required

Organization administrators must enable MCP access in Settings > Org > Overview
Once enabled by an admin, all Mixpanel users in the organization can use the MCP integration
The MCP toggle controls access for the entire organization

User Access ScopeAfter the organization admin enables MCP:

Any Mixpanel user with any role can connect to the MCP integration
Users can only query data from projects they already have access to within Mixpanel
Existing Mixpanel permissions and project access controls remain in effect

Recommended Implementation Path

Organization admin enables “Enable MCP” toggle in organization settings
Individual users complete the integration setup following the Implementation Guide
Users authenticate with their Mixpanel account during the connection process
Once authenticated, start by asking: “What Mixpanel projects do I have access to?”
Select a project from the response to focus your analysis
Ask any questions about your Mixpanel data—refer to Use Case Examples below if you need ideas
Begin with test data to validate functionality
Conduct security review before production data access
Train teams on natural language query techniques
Expand usage across product and analytics teams

Implementation Guide

Prerequisites (for Claude Free, Cursor & Gemini CLI)

Install Node.js (if needed):

Option 1: Claude Pro

Setup Steps:

Navigate to claude.ai or open Claude desktop app
Go to Settings → Connectors → Add Custom Connector
Configure the integration:
- Integration Name: Mixpanel (or preferred name)
- Integration URL
  * US:
  * EU:
  * IN:
Click Connect and complete Mixpanel authorization

For additional information on setting up Claude connectors, see Claude’s connector documentation.

Option 2: Claude Free

Additional Requirements:

Claude desktop application
Command line access

Setup Steps:

Open Claude desktop app
Navigate to Settings (from window nav bar) → Developer → Edit Config
Add the following configuration:
Run authorization command in terminal:

For EU,

For IN,

Complete Mixpanel authorization via the provided link

Option 3: Cursor

Additional Requirements:

Cursor application
Command line access

Setup Steps:

Open Cursor application
Navigate to Cursor Settings → MCP & Integrations → MCP Tools → New MCP Server
This opens MCP.json where you paste the following configuration:
Run authorization command in terminal:

For EU,

For IN,

Complete Mixpanel authorization via the provided link

Option 4: Google Gemini CLI

Additional Requirements:

Gemini CLI installed
Command line access

Setup Steps:

Install the Gemini CLI by following Gemini CLI installation instructions.
Authenticate to Google Gemini using your Google account:
Configure Mixpanel MCP as a remote server and authorize:
For EU,
For IN,
Running this command will provide a link in your terminal to authorize your access with Mixpanel. Follow the authorization steps in your browser. After logging in via OAuth, you should see a confirmation message.
Verify Configuration (or Create It Manually):
After authorization, your file for the Gemini CLI should be updated automatically. You can find this file at . To check, run:
The contents should be similar to the following:
This configuration ensures that the Gemini CLI knows how to connect to the Mixpanel remote server. If is not available, create it manually using the same contents as above.
Complete Mixpanel authorization via the provided link (if not already done).
Run the Gemini CLI Once configured, you can use the Gemini CLI with the specified model by running this command:

Option 5: ChatGPT (OpenAI)

Additional Requirements:

OpenAI/ChatGPT account with MCP access

Setup Steps:

Enable Developer Mode
- Go to Settings → Connectors → Advanced
- Enable “Developer Mode” (see https://platform.openai.com/docs/guides/developer-mode)
Add the Mixpanel MCP connector
- Create a new connector/integration in the ChatGPT Connectors UI
- Provide these values:
  * Name: Mixpanel MCP
  * Description: The MCP integration provides access to all core Mixpanel analysis capabilities through natural language queries.
  * MCP Server URL
  * US:
  * EU:
  * IN:
  * Authentication type: OAuth
- Click Create and then connect to Mixpanel and follow the OAuth flow to authorize Mixpanel access.
Publish connector to workspace (optional)
- To make the connector available to other members: go to Settings → Apps & Connectors → Mixpanel → Publish

Use Case Examples

Channel Performance Analysis

Query example: “Which channels drive the most users?”
Business Value: Identify top-performing acquisition channels for budget allocation

Retention Intelligence

Query: “Which channels drive users who retain best?”
Business Value: Understand quality vs. quantity in user acquisition strategy

Conversion Optimization

Query: “Show me signup conversion rates by source”
Business Value: Optimize funnel performance by traffic source

User Journey Analysis

Query: “What’s the typical user journey for power users?”
Business Value: Identify patterns in high-value user behavior for product development

Advanced Analytics

Query: “Compare user behavior across different feature roll outs”
Business Value: Measure feature impact and optimize product decisions

User Feedback Analysis

Query: “Why did this user submit negative feedback? Analyze their session replays.”
Business Value: Understand user context and behavior patterns by analyzing replay footage in combination with your event data.

Usage Analysis

Query: “What features does this user use most? Analyze their session replays”
Business Value: Better understand specific customers usage patterns without needing to build reports or watch dozens of replays

Troubleshooting User Issues

Query: “This user reported an issue; what happened in their replays?”
Business Value: Accelerate support investigations by getting instant answers that combine events and replay context

Key Benefits

Democratized Analytics

Natural language queries eliminate technical barriers
Instant insights without dashboard complexity
Accessible to non-technical stakeholders

Enhanced Productivity

Reduce time-to-insight from minutes to seconds
Eliminate context switching between tools
Enable real-time decision making
Save hours daily by analyzing replays and events together without manually piecing together

Strategic Value

Accelerate product discovery processes
Enable data-driven conversations across teams
Improve accessibility of product intelligence

Security & Compliance Considerations

Data Handling Important: This integration enables querying your Mixpanel data through AI services (Claude, Cursor, etc.). Review your organization’s data handling policies before implementation.

Access Controls

Organization administrators must enable MCP in Settings > Org > Overview
Users can only query data from projects they already have access to within Mixpanel
Existing Mixpanel permissions and project access controls remain in effect

Best Practices

Start with demo or test datasets
Conduct internal security review before production use
Consider compliance requirements (GDPR, CCPA, etc.)
Establish usage guidelines for teams
Monitor data sharing patterns

Technical Specifications

Supported Platforms

Claude Pro: Web interface with native MCP support
Claude Free: Desktop app with configuration setup
Cursor: AI-powered code editor integration

Integration Architecture

Protocol: Model Context Protocol (MCP)
Connection: Server-sent events (SSE)
Authentication: OAuth with Mixpanel
Endpoint:

Troubleshooting

Common Issues

Node.js not found: Install using
Authorization fails: Ensure proper Mixpanel account permissions
Desktop app issues: Restart application after configuration changes
Free user limitations: Remote URL integrations only supported in desktop app

Computed Properties Revenue Analytics