Start here
Pick a project
Find something you pay for and build a replacement
The strongest CloudFALL submissions will start with something specific: a service, a price, a month. For example "The $17.99 I pay Netflix, but I have dvds at home" is a project waiting to happen.
Things people rent, and what runs instead#
| You pay for | Runs instead | Port | What you actually get |
|---|---|---|---|
| Netflix | Jellyfin | 8096 | Rip the discs you own, stream them to every screen in the house like netflix. |
| Spotify | Navidrome | 4533 | High quality music for free, with no ads that you own forever. |
| Google Photos | Immich | 2283 | Automated backups for all of your photos but you don't owe google $25/M for the rest of time. |
| iCloud+ | Syncthing | 8384 | Every photo and backup your phone makes on every device you own |
That table is a starting point. The test is in what reviewers check: does it replace something proprietary(privately owned) that people pay for, and is it open source when you're done.
Ways to make it yours#
Installing Jellyfin isn't a project!!! somebody already built Jellyfin! Building something around the problem is:
- The thing that's missing from an existing tool. A phone app for the service you run. A migration tool that gets your data out of the paid thing.
- The glue. Something that watches a folder, tags what lands in it, and files it where three other services can find it.
- The hardware end. An antenna that pulls in radio and streams it to your network. A sensor that logs something the paid app charges you to see.
- The subscription nobody has replaced yet. Gym trackers, recipe apps, language courses, parking apps. Look at the last three things that charged your card.
Hardware projects#
Hardware is in scope as long as it's useful to a homelab or on-prem setup, even somewhat indirectly the antenna example above is the canonical one. The same rules apply: it has to be documented well enough that someone else could build it, and open source means the schematics and firmware too, not just a photo of the finished box, that also means tracking your time using lapse
Scoping it#
Aim for something that runs end to end, then make it better. A working thing with three features beats a beautiful thing that doesn't start. You can submit muliple times even if you change something and add a new feature so don't be afraid to ship.