PostHog parity shipped

AI-native analytics that ships the action, not just the chart

TrackCrumb tracks user behavior, surfaces AI insights, and lets you act in one click — without stitching together five tools.

TrackCrumb · Dashboard

DAU

12,483

+8.3%

MRR

$48,290

+12.1%

Churn rate

2.1%

−0.4%

Daily Active Users — last 7 days

MonTueWedThuFriSatSun
✨AI detected a 38% DAU spike Saturday — likely correlated with your Product Hunt launch.

Features

Everything you need to build a better product

  • Lightweight SDK

    ≤25 KB gzipped. Autocapture clicks, pageviews, form submits. First event in 5 min.

    // Install
    npm install @trackcrumb/sdk
    // Init
    TrackCrumb.init("your-write-key")

  • Trends, funnels, retention

    Time-series, ordered conversions, weekly cohort grids — no SQL required.

  • Session replay & heatmaps

    rrweb v2 playback with PII masking by default. Click density per URL.

  • AI-Native Intelligence

    Multi-turn AI chart builder, anomaly detection with natural-language explanations, predictive churn scoring — gated by per-tenant budgets and zero-retention LLM routing.

    AI chart builderAnomaly detectionChurn predictionAI explanations

  • Cohorts & groups

    Dynamic + static segments with compound conditions. Company-level analytics across every chart.

  • In-app messages

    Modals, banners, tooltips. NPS + rating + choice + text. Rule builder with cohort targeting.

  • Experiments & flags

    Bayesian engine, SRM auto-pause, holdouts, segment breakdown, power calculator.

AI-native

Not a chatbot bolted onto a dashboard

The model has structured access to your schema, your metrics, and your users — gated by per-tenant budgets and zero-retention LLM routing.

  • AI chart builder

    Multi-turn chat with conversation history; Claude writes ClickHouse SQL and renders the chart.

  • Anomaly detection

    Rolling 28-day median + MAD on your top 20 metrics; flagged automatically.

  • AI anomaly explanations

    Claude Haiku explains why a metric moved, with semantic-cache for repeat queries.

  • Churn prediction

    XGBoost binary classifier + Cox PH survival curves, refreshed nightly with SHAP attribution per user.

  • AI "why" for at-risk users

    Natural-language summary of churn drivers for each high-risk account.

At-risk users

Alice M.
87%
Bob K.
74%
Carol S.
61%
David L.
45%

Predictive Intelligence

XGBoost binary classifier + Cox PH survival curves, refreshed nightly. SHAP attribution explains exactly why each user is at risk — so you can intervene before they churn.

✨ AI: Alice M. is high-risk — dropped from daily to weekly sessions, skipped onboarding step 3.

Why TrackCrumb

How we compare

Factual feature comparison. Claims sourced from public docs.

FeatureTrackCrumbPostHogMixpanelAmplitude
AI-powered insights (built-in)✅Partial❌Partial
Predictive churn scoring✅❌❌✅ (Growth+)
Session replay✅✅❌❌
Feature flags & A/B tests✅✅❌Partial
In-app messaging✅❌PartialPartial
SDK gzip size≤25 KB~90 KB~70 KB~80 KB
Self-hostable (Docker)✅✅❌❌
AI actions (act on insight)✅❌❌❌
AI anomaly explanations✅❌❌❌
Web Vitals built-in✅Partial❌❌
Error tracking + sourcemaps✅✅❌❌
Group (company-level) analytics✅✅PartialPartial

Early customer testimonials will appear here.

Pricing

Simple, event-based pricing

Start free, upgrade as you grow. No overage surprises.

Loading plans…

FAQ

Frequently asked questions

Ready to see your users clearly?

Create your free workspace in under a minute. No credit card required.

Already have an account? Log in.