Which self-hosted projects are actually still alive?

Every directory tells you what exists. None tell you what is still maintained. We score 1352 self-hosted projects from 0 to 100 on real repository activity — commit volume, release cadence and recent trend — so you can rule out the dead ones before you spend a weekend deploying them.

1,352
projects scored
84
categories
100
SaaS migrations
778
head-to-heads
2026-10-01
data as of

Leaving a SaaS product?

The self-hosted replacements for the tools people most often want to drop, ranked by health.


Best maintained right now

Highest health scores across the whole catalogue, recomputed on every rebuild.



Browse by category


The State of Self-Hosted Software

We aggregate every score into one picture of the ecosystem: which categories are thriving, which runtimes age worst, and how much of the catalogue has quietly stopped moving. Recomputed on every rebuild, free to cite.

Read the report →

Why a health score

Self-hosting fails in a specific way. You pick a project because it looks good and has stars, you spend a weekend deploying it, and eighteen months later you discover the last commit was before you started. Migrating away costs more than the original setup did.

Star counts do not tell you this — they are cumulative and never go down. A project with 30,000 stars and no commits in a year looks more attractive than a 400-star project shipping weekly, and it is the worse choice. The score on this site is designed to invert that, by ranking every project against the whole catalogue on activity that actually decays. See exactly how it is computed.