I wanted analytics for a side project but didn't want to hand my users' data to a third party, and I didn't want to pay for PostHog's cloud plan. So I built InsightFlow — a self-hosted product analytics platform with an AI layer that actually tells you what the data means .
The difference from standard analytics is the AI insights feature. Most analytics dashboards show you numbers and leave the interpretation to you. InsightFlow looks at your data and tells you things like "users who land on the pricing page from Google have a 3x higher conversion rate than those from social media" .
🛠️ What it does:
Event Tracking — Drop a single JavaScript snippet into your site and it starts capturing pageviews, clicks, form submissions, and custom events.
User Analytics — Unique visitors, session counts, pageview trends over time.
Funnel Analysis — Define a series of steps and see where users drop off, with conversion rates at each step.
Retention Tables — Cohort-based retention analysis to see if your product actually has stickiness.
Click Heatmaps — Visual overlay showing where users click on each page .
AI Insights — Runs across your data every day and generates observations and recommendations .
Anomaly Detection — Flags unusual patterns in your metrics automatically.
🧰 Tech Stack
Python, FastAPI, React, TypeScript, PostgreSQL, Redis, and optional Ollama for local AI.
🔗 Get Started
Check out the repo: github.com/isidhartha/insightflow-ai
Top comments (0)