About
Product analytics you run yourself.
Lyraflow records what people do in your product, ties an anonymous visitor to the account they eventually create, and lets you ask who did what. It runs under Docker on your own machine, and the events never leave it.
What it is, said against something you already know
Lyraflow is a self-hosted alternative to Mixpanel, Amplitude and Heap. It is not an alternative to Google Analytics, and the difference is not a positioning nicety — it is the whole shape of the product.
Web analytics counts visits: how many people came, from where, to which pages. There are a dozen good self-hosted tools for that and most of them are easier to run than this one. Product analytics answers a different question — which people did what, in what order, and did they come back — and it cannot be answered without deciding that the anonymous visitor from two weeks ago and the account that signed up on Tuesday are one person.
That decision is the part Lyraflow is built around. If two people share a device, each event is attributed to whoever was signed in at that moment, not to whoever used it last. Every screen reads the same stitched identity, so a person who looks like two people in a funnel looks like two people on their own profile, where you can actually see why.
Why self-hosted is the point and not a feature
Product analytics means recording, in detail, what your customers do. Sending that to someone else's servers is a decision most teams make once, quietly, by picking a tool. Lyraflow exists so the decision does not have to be made: events go to ClickHouse and identity to Postgres, one container each, on infrastructure you already have.
The practical consequences are the ones worth stating. There is no data-processing agreement to sign, because there is no processor. There is no per-event pricing to model, because nobody is metering you. And there is nobody to ask for an export, because you are already holding the database.
How far along it actually is
v0.15, and early. The first commit was on 1 August 2026 and the first tagged release two weeks after it. Every minor version between that one and this has been a real release rather than a number. As far as we know, nobody outside the project is running it in production yet.
What exists is a complete ingest and identity layer, segments, funnels, retention grids, trends, saved reports, dashboards, shared dashboard links, a person profile, deletion and export, event retention, per-project quotas, and backup and restore — with a web UI over all of it. What does not exist is also written down rather than left to be discovered: there are no journeys and no path analysis, no alerting, no scheduled export and no email digest of any kind, and neither a funnel nor a retention grid can be broken down by a dimension the way a trend can.
Every screen in the documentation ends with what that screen does not do yet, and every limit in it has a number attached. That is the habit this project is most careful about: a tool you are trusting with your customers' behaviour should tell you where it stops before you find out.
The licence, in plain terms
Lyraflow is fair-code under the Sustainable Use License. The source is always visible. You may self-host it, use it and modify it for your own business, including commercially, and you do not owe anybody anything for doing so.
It is not an OSI-approved open source licence, and calling it open source would be a factual error about the licence rather than a stylistic slip. The practical difference is one sentence long: you may not sell Lyraflow as a hosted service to other people. That is the line that keeps a hosted version viable later, which is what pays for the self-hosted one staying free.
Code contributions need a signed CLA, and the CLA is not in force yet — the text is drafted and the signing check is committed, but both are switched off until it has had a legal review. Until then, please open an issue rather than sending a pull request. Documentation fixes, bug reports and issues are welcome now and need no agreement of any kind.
Who builds it
Lyraflow is built by Cem Hurturk. One person, which is the fact that explains most of the others: the roadmap is short, replies take a few days, and the public issue tracker is the honest version of what is planned. There is no support rota behind it and this page will not pretend otherwise.
What there is no price for
There is no pricing page, because there is nothing to buy. Lyraflow is software you run yourself, and there is no hosted plan to compare it against. Any page or file quoting a Lyraflow price did not come from Lyraflow.
Where to go next
Getting started is four steps and about five minutes, most of it waiting for Docker. The documentation is the whole surface. The source is on GitHub, and if you would rather ask a person, contact lists the ways.