Free Software, Proprietary Pipeline

This talk is for anyone who has ever looked at their Android app's CI pipeline and just... accepted it.

I self-host almost everything. Gitea, Traefik, Portainer, Uptime Kuma. So when I started exploring open source Android development, I assumed the pipeline story would be the same: find the FOSS tool, self-host it, move on with life. Spoiler: it was not that simple.

Most FOSS Android apps quietly depend on proprietary services at every single stage. Not one stage. Every stage. I audited each one and asked the same question at every step: why are we using proprietary tools here, and what happens when you actually try to replace them?

The four stages I investigated:

• Code hosting and CI triggering, where GitHub has become the unquestioned default

• Building the APK, where Codemagic and Bitrise run on servers you do not control

• Signing the APK, where Google Play Signing quietly takes ownership of your keystore

• Distribution, where Firebase and the Play Store are treated as the only options

For each stage I tried replacing the proprietary tool with something self-hostable. Forgejo Actions, self-hosted runners, Fastlane, F-Droid. This talk is the honest story of what worked, what did not, and why I had to read three wiki pages to do something that takes five minutes on Codemagic.

• Why FOSS mobile development quietly defaulted to proprietary pipelines at every stage and what that actually costs us in terms of software freedom

• A stage by stage breakdown of what a fully self-hostable Android CI pipeline looks like in 2026 and where each tool falls short

• Where the real gaps are, missing documentation, broken tooling, and problems nobody has bothered to fix yet

• What the IndiaFOSS community could realistically build or fix to make open source Android development truly open source all the way down