Compare
An Amplitude alternative you run on your own server.
Amplitude is a hosted service, run from data centres in the US and the EU. Lyraflow does the same kind of work on a machine you run, so the events never go anywhere else. It is also much younger and much smaller, and this page says where.
Every claim about Amplitude on this page links Amplitude's own documentation, and each was checked on 22 September 2026. If something here has gone out of date, tell us and it will be corrected.
The short version
| Amplitude | Lyraflow | |
|---|---|---|
| Where it runs | Data centres hosted by AWS. Its documentation describes no self-hosted option | Your server, under Docker. There is no hosted version |
| Where the data is kept | The US, or the EU in Frankfurt | Wherever you put the server |
| Getting raw events out | An export API, up to 4 GB and 365 days per request, with S3 export beyond that | Already in your own ClickHouse database |
| Changing a user id | Not possible once it is set | An alias call merges the old id into the new one, permanently |
Where your data lives
Amplitude's security and privacy page says it maintains data centres hosted by AWS in the US and in the EU, and that it chose Frankfurt for the EU one. Amplitude launched the EU data centre in 2021. The same page says an organisation provisioned in the EU keeps its data in the EU when it uses Amplitude's AI features. None of the Amplitude pages checked for this comparison describe a self-hosted or on-premises deployment.
Lyraflow has no regions because it has no servers of its own. Events go to ClickHouse and identity to Postgres, both in containers on the machine you install it on, and nothing is sent anywhere else. There is no data-processing agreement to sign with Lyraflow, because Lyraflow never receives the data.
Getting your data out
Amplitude's Export API returns every event in a time range as a zipped archive of JSON files, one event per line. Its documented limits are 4 GB per request, a 365-day window per request, and up to two hours between an event arriving and it being available to export. For anything larger, Amplitude's documentation points to export to Amazon S3.
With Lyraflow there is nothing to ask for, because the database is already yours: you can query ClickHouse directly. Over the API, events can be read a page at a time, and one person's complete record exports as NDJSON for a subject-access request. There is no bulk export endpoint, no scheduled export and no sync to a warehouse or to other tools.
Identity
Amplitude counts a user by combining three identifiers into one profile: a device id from its SDK, a user id you set after login, and an Amplitude ID it computes to connect the two. Its documentation says a user id cannot be changed once it is set. It also notes that for data in Redshift, which is immutable, an anonymous event keeps its original Amplitude ID in the raw data even after the merge is recognised, so the raw data and the merged view in the product can differ.
Lyraflow binds a device to a person with an identify call, and every
event that device sent before it, as well as after, resolves to that person. When
a device is later bound to someone else, as on a shared computer, the history
splits at that moment: each event belongs to whoever was signed in when it
happened. When a user id changes, an alias call merges the old id
into the new one. That merge cannot be undone.
Identity resolution has the details.
What Amplitude does that Lyraflow does not
- It runs itself. Amplitude is a hosted service. Lyraflow is a server you run: HTTPS, backups, upgrades and disk space are yours to look after. If nobody on your team wants that job, a hosted tool is the better choice.
- It exports at scale to your own storage. Amplitude documents export to Amazon S3 for volumes beyond its API. Lyraflow has no bulk export and no scheduled export; its data is already on your disk, but moving it anywhere else is your own query.
- It already has your history. Amplitude keeps what you have already sent it. Lyraflow has no importer, and its ingest API clamps every timestamp to within 24 hours of arrival, so months of past events cannot be sent to it over HTTP. Moving means starting from the first event Lyraflow receives.
And Lyraflow's own gaps. There is no session replay, no feature flags, no experiments and no surveys. There is no path analysis, no alerting, no scheduled export and no email digest. A funnel and a retention grid cannot be broken down by a dimension the way a trend can. There is one admin account and no roles. Ingest never fills country, region or city yet, because there is no GeoIP lookup (#12); the countries in the demo are generated seed data. There is no published container image, so the install builds one from the checkout. Lyraflow's first commit was on 1 August 2026, it is built by one person, and as far as we know nobody outside the project runs it in production yet.
Who it suits
Lyraflow suits a team that wants funnels, retention, trends and segments over its own users, would rather the events never left its own infrastructure, and can run three containers. It does not suit a team that wants a hosted service, needs its historical data carried over, or needs the product to be further along than Lyraflow is today. The live demo is the quickest way to tell which you are, and getting started takes about five minutes.
Sources
Every Amplitude page this comparison was checked against, on 22 September 2026:
- https://amplitude.com/security-and-privacy
- https://amplitude.com/blog/understanding-privacy-and-analytics
- https://amplitude.com/docs/apis/analytics/export
- https://amplitude.com/docs/data/csv-import-export
- https://amplitude.com/docs/data/sources/instrument-track-unique-users
- https://amplitude.com/docs/get-started/identify-users
There is a separate page comparing Lyraflow with PostHog, which can also be self-hosted.