Upfront: I did not build this. It is an open-source project on GitHub (retentioneering/retentioneering-tools, Apache-2.0). It is listed here on star count, maintenance activity and how well it fits day-to-day storefront operations.
What it does
- Path graphs — renders event streams as transition graphs, revealing that a lot of users take routes you never designed — something a standard funnel cannot show.
- Funnels and retention — custom funnel steps and retention windows, with cohort comparison across user segments.
- MCP server included — connect it to an AI session and have the agent run the analysis instead of writing all the pandas yourself.
- Reproducible — analyses live as notebooks or scripts, so conclusions can be audited and re-run rather than being one-off exploration.
How to use it
Install with pip; it is usually driven from Jupyter:
pip install retentioneering
You need an event stream — user ID, event name and timestamp is enough — exported from GA4 or queried from your database. The docs ship a sample dataset to work through first.
When to use it, and when not to
A steeper learning curve than the others, and some pandas is assumed. For a ready-made funnel UI use PostHog; the value here is the custom path analysis standard tools cannot do.


