Mixpanel

@mixpanel what were our top 5 events yesterday and how do their counts compare to the day before?

Based on yesterday's data, your top 5 events were: 1) page_view (12,847 events, +8.3% vs prior day), 2) button_click (4,231 events, -2.1%), 3) form_submit (1,892 events, +15.7%), 4) video_play (1,203 events, +4.2%), and 5) search_query (987 events, -6.4%). The form_submit event showed the strongest growth, while search_query declined notably. This data reflects your production project for the 24-hour period ending midnight UTC.

This example uses the Get Today's Top Events tool to surface high-level engagement trends. Be aware of the 60 queries/hour rate limit — repeated requests in a conversation can exhaust your quota. The percentage changes help identify anomalies worth investigating further.

@mixpanel show me the event activity for user ID abc-123 over the last 7 days, focusing on purchase-related events

User abc-123 triggered 23 events in the past week. Purchase-related activity: viewed product_detail 8 times (most recently 2 hours ago for item SKU-9921), added 3 items to cart, initiated checkout twice, and completed 1 purchase (order #ORD-4482, $127.50, 3 days ago). Non-purchase events included 6 page_view events and 4 search_query events. The user appears to be in an active browsing phase with one recent conversion.

This example demonstrates the Get Profile Event Activity tool, which returns a time-ordered feed of user actions. Requires read access to user profiles in your Mixpanel project. Useful for debugging individual user journeys or investigating support tickets tied to specific accounts.

@mixpanel run a JQL query to calculate the average session duration for users who completed a purchase in the last 30 days, grouped by traffic source

Executed custom JQL query against 30 days of event data. Results: Organic search users averaged 8.4 minutes per session before purchase (n=1,247 conversions). Paid social users averaged 5.2 minutes (n=892). Direct traffic users averaged 11.7 minutes (n=634). Email campaign users averaged 6.1 minutes (n=423). Query processed 4.2 GB of data in 47 seconds. This suggests organic and direct users exhibit more exploratory behavior pre-purchase compared to paid channels.

This showcases the Execute JQL Query tool for custom analytics beyond Mixpanel's standard reports. Requires familiarity with JQL syntax. Subject to strict limits: 60 queries/hour, 5 concurrent max, 2-minute timeout, 5 GB data cap. Complex queries can hit these ceilings quickly — test incrementally and cache results when possible.

When Mixpanel MCP makes sense for small product teams

A 5-person product team ships features every sprint and needs to review adoption metrics in their Friday retro. The Mixpanel MCP is a good fit here because the team can query event counts and property breakdowns directly in Switchy without context-switching to the Mixpanel dashboard. The 60-query-per-hour limit is fine for a weekly cadence—you're pulling 8-12 queries per session, not hundreds. The JQL tool lets you write custom queries for edge cases without waiting on a data analyst. If your team reviews metrics daily or runs more than 10 concurrent analyses, you'll hit rate limits and need to batch requests or move to the Mixpanel UI. For weekly reviews with moderate query volume, this MCP saves 15 minutes of tab-switching per meeting.

Using Mixpanel MCP to debug user drop-off patterns

A 3-person customer success team notices a cohort of trial users churning at day 7 and needs to understand which features they skipped. The Mixpanel MCP works well here because you can pull profile event activity for specific users and compare their behavior to successful cohorts—all in one Switchy thread. The Get Profile Event Activity tool surfaces the exact event sequence for each user, and the Get Top Event Properties tool shows which feature interactions correlate with retention. The 2-minute query timeout is rarely an issue unless you're analyzing profiles with 50k+ events. If you're investigating more than 60 users per hour, you'll need to stagger your queries or export a CSV from Mixpanel first. For ad-hoc churn investigations with 10-30 user profiles, this MCP turns a 2-hour spreadsheet exercise into a 20-minute conversation.

When Mixpanel MCP hits limits for high-frequency dashboards

A 6-person growth team runs daily standups where they review yesterday's signup funnel, activation rate, and top-performing campaigns. The Mixpanel MCP is borderline here. The Get Today's Top Events tool gives you a quick snapshot of what changed overnight, and the Get Aggregate Event Counts tool pulls funnel metrics without opening a browser. But if your team runs 15+ queries per standup and does this 5 days a week, you're burning through the 60-query-per-hour limit by Tuesday. The 5-concurrent-query cap also means you can't parallelize requests during live meetings. If your daily review is under 10 queries and you're okay with sequential execution, the MCP keeps your standup in Switchy. If you need real-time dashboards or 20+ daily queries, build a Slack bot that hits Mixpanel's API directly or stick with the Mixpanel UI.