---

name: mp_mcp description: Use the Mixpanel MCP server for analytics. Triggers on mentions of Mixpanel MCP tools, MCP resources, analytics queries via MCP, segmentation, funnels, retention, cohort_comparison, product_health_dashboard, ask_mixpanel, diagnose_metric_drop, guided_analysis, fetch_events, or SQL queries on local DuckDB data.

Mixpanel MCP Server

The Mixpanel MCP server exposes analytics capabilities through 40+ tools, 8 resources, and 8 prompts.

Reference Files

When you need detailed information, read these reference files:

File	When to Read
tools.md	Complete tool signatures, parameters, return types for all 40+ tools
resources.md	Resource URIs (workspace://, schema://, analysis://, users://, recipes://)
prompts.md	Guided workflow prompts (analytics_workflow, funnel_analysis, retention_analysis, etc.)
patterns.md	Usage patterns, JSON property access, error handling, best practices

Tool Categories

Category	Purpose	Key Tools
Discovery	Explore schema	list_events, list_properties, list_funnels, list_cohorts, top_events
Live Query	Real-time API	segmentation, funnel, retention, jql, event_counts, property_counts
Fetch	Download data	fetch_events, fetch_profiles, stream_events, stream_profiles
Local	SQL analysis	sql, sql_scalar, list_tables, sample, summarize, event_breakdown
Composed	Multi-primitive	cohort_comparison, product_health_dashboard, gqm_investigation
Intelligent	AI-powered	ask_mixpanel, diagnose_metric_drop, funnel_optimization_report
Interactive	Guided workflows	guided_analysis, safe_large_fetch

Quick Patterns

Explore -> Query -> Analyze

1. list_events() -> top_events()     # Find what to analyze
2. list_properties(event)            # Understand the data
3. segmentation() or funnel()        # Quick insights
4. fetch_events() + sql()            # Deep local analysis

Diagnostic Investigation

1. diagnose_metric_drop(event, date)  # Automatic root cause analysis
2. cohort_comparison(a_filter, b_filter)  # Isolate segment differences
3. ask_mixpanel("Why did X happen?")  # Natural language queries

Product Health

1. product_health_dashboard()         # AARRR snapshot
2. guided_analysis(focus_area)        # Interactive drill-down
3. funnel_optimization_report(id)     # Bottleneck analysis

Resources (Read-Only)

URI	Content
workspace://info	Project connection status, tables
schema://events	All tracked event names
schema://funnels	Saved funnel definitions
schema://cohorts	Saved cohort definitions
analysis://retention/{event}/weekly	12-week retention curve
analysis://trends/{event}/{days}	Daily event counts
users://{id}/journey	User's 90-day activity
recipes://weekly-review	Weekly analytics checklist

JSON Property Access (Critical)

Events and profiles store properties as JSON. Access with DuckDB JSON syntax:

-- DuckDB SQL (local queries)
SELECT properties->>'$.country' as country FROM events
WHERE properties->>'$.plan' = 'premium'

Filter Expressions

Use SQL-like syntax for filtering in API calls:

# WHERE parameter
where='properties["amount"] > 100 and properties["plan"] in ["premium", "enterprise"]'

# ON parameter (segmentation)
on='properties["country"]'

Error Handling

All tools return structured errors with actionable suggestions:

Error	Cause	Solution
TableExistsError	Table already exists	Use append=True or drop table first
AuthenticationError	Invalid credentials	Check workspace_info()
RateLimitError	API rate limited	Wait and retry

Middleware

The server includes automatic:

• Rate limiting: Respects Mixpanel API limits (60 req/hour for query, 3 req/sec for export)
• Response caching: Discovery operations cached for 5 minutes
• Audit logging: All requests logged with timing