I'm irked how "Firebase alternative" has come to mean "hosted database with an API layer". I mean, the key pitch of Firebase has always been that it's a realtime database. With both Supabase and Trailbase this seems a bit of a bonus aspect, kinda sorta half supported for a few things here and there. In Firebase meanwhile, as well as its successor Firestore, everything can be monitored in realtime. This is its key feature and all design choices derive from that. Notably all its weird limitations of the schema possibilities and query options derive from the goal to make a simple database that's 100% realtime-observable.
It's just super weird to me to see something presented as a "Firebase alternative" without a single realtime data sync example on the homepage. Trailbase has one deep, deep down the docs and nothing is explained about the "subscribe" limitations. That's not a Firebase alternative, it's just a hosted database. That's like calling a Chromebook an "iPhone alternative" because it lets you access WhatsApp and Twitter just like an iPhone would.
ps. I haven't followed Supabase a lot so maybe they got on top of this since, but it was pretty disappointing to me back when they launched to much fanfare. Like the only realtime thing they had was a counter or something like that.
We likely agree that firebase is superb. But your criticism of using ‘alternative’ is unjust both in terms of the breadth of firebase, and why a competitor might target someone about to choose Firebase.
skrebbel · 32d ago
Actually I think I'm just old. To me "Firebase" still means the "Firebase Realtime Database" but after they got acquired by Google, they took the Firebase brand and started building all kinds of features that would let non-backend engineers ship apps without needing a backend, including an improved version of the realtime database (Firestore).
Somehow I've always read "Firebase alternative" as "alternative to Firebase Realtime Database" but now I see that what Supabase & co mean is "stuff you can put on a server once and then you won't need to do any backend stuff anymore", ie the current positioning of the Firebase brand.
I stand corrected.
(sidenote, Google's Firebase acquisition has to be one of the best-executed devtool acquisitions right? Nobody would've been surprised if they had just gobbled Firebase up, spread the team across GCP, put the product on life support and after a year told customers to just move to GCP proper, somehow. Instead they realized that "GCP Easy Mode" was worth having a separate brand for and they executed excellently on that)
creativenolo · 32d ago
Yes - totally agree on side note. Took me a while to realise that firebase is now a high-level hat sat on top of GCP that you can take off, and put back on, as you please. Very neat.
It’s insanely lean to get started. But at the end of the day, its GCP underneath with GCP billing (i.e. no spend caps). The easy mode UX, and the risk of real expenditure, sit in contradiction to each other. Of course, there a very generous free tier to placate that. But things like realtime need to be used with real cost considerations, and not frivolously once you need a feature only available on pay as you go.
trailbase-alt · 33d ago
I think it's fair criticism and certainly an area where TrailBase can improve (it's still young). FWIW, I used to avoid the "Firebase" comparison and be more descriptive with a long list of features. That confused a lot of folks. Firebase-like seemed to get the rough product-ballpark across more effectively. If I had to guess, I'd say Supabase probably arrived at a similar conclusion.
Either way, they are all rather broad products and many users only use subsets.
Thanks for speaking up, appreciated.
skrebbel · 32d ago
Fwiw I changed my mind and I now think that your positioning is fine (see adjacent comment). Don't make it harder for people to grok at the behest of HN purists like me! :-)
oulipo · 33d ago
Trying the online demo but it seems that it doesn't work properly, getting a lot of errors
trailbase-alt · 33d ago
Sorry to hear. The beauty of giving folks admin access and that's why we can't have nice things :). Excuses aside, in case you're still interested maybe try a bit later. The demo resets regularly. If you still run into issue, please don't hesitate to send a quick, low-key bug report: https://github.com/trailbaseio/trailbase/issues . Strong wording welcome
CoolCold · 31d ago
sounds like a nice side project for myself / something new (wth is that hype on Firebase and alternatives?) to look into on weekend like create per-user (likely with some limit of total 10 or so instance) demo env - average sysadmin like me should be able to handle it.
If i'll do something, will let you know if you wish
barfoos · 30d ago
> wth is that hype on Firebase and alternatives?
It's mostly about communicating the intend and a testament to Firebase establishing this rather broad product category, i.e. DB + dashboard + auth + files + ... You could probably also call it "a bucket of generally useful stuff", just doesn't have the same ring :).
> create per-user (likely with some limit of total 10 or so instance) demo env - average sysadmin like me should be able to handle it.
Totally feasible. There are also kubernetes deployment manifests in the repo. That said, I don't mind the shared instance too much. There's pros and cons. Happy to chat more.
> If i'll do something, will let you know if you wish
Honestly they spent more time comparing supabase to pocketbase
trailbase-alt · 33d ago
The Supabase comparison is pretty poor and needs improvement. In many ways TrailBase and PocketBase are architecturally a lot similar. The key-difference of both compared to Supabase is that they're a single executable, easy to self-host, less resource hungry and in many cases faster (for better or worse due to their simpler architecture). In case you're interested: https://trailbase.io/reference/benchmarks/
laveur · 33d ago
This is pretty cool. Just needs clients for iOS and Android and it would be a rock solid replacement for a lot of large enterprises, or privacy minded companies.
trailbase-alt · 33d ago
Privacy, data ownership and reduced reliance on public clouds were definitely motivations.
A Java/Kotlin client is certainly on the short-list. We recently added one for Swift. There are also clients for Dart, TypeScript, and C# in case you're more into cross-platform app development (Flutter, React Native, and Maui/Avalonia respectively)
aklemm · 33d ago
How did you evaluate it so thoroughly?
songtianlun1 · 33d ago
good! i will try it in my next project.
trailbase-alt · 33d ago
Appreciated. If you run into any issues, please don't hesitate to file a quick report. Strong wording welcome :)
It's just super weird to me to see something presented as a "Firebase alternative" without a single realtime data sync example on the homepage. Trailbase has one deep, deep down the docs and nothing is explained about the "subscribe" limitations. That's not a Firebase alternative, it's just a hosted database. That's like calling a Chromebook an "iPhone alternative" because it lets you access WhatsApp and Twitter just like an iPhone would.
ps. I haven't followed Supabase a lot so maybe they got on top of this since, but it was pretty disappointing to me back when they launched to much fanfare. Like the only realtime thing they had was a counter or something like that.
We likely agree that firebase is superb. But your criticism of using ‘alternative’ is unjust both in terms of the breadth of firebase, and why a competitor might target someone about to choose Firebase.
Somehow I've always read "Firebase alternative" as "alternative to Firebase Realtime Database" but now I see that what Supabase & co mean is "stuff you can put on a server once and then you won't need to do any backend stuff anymore", ie the current positioning of the Firebase brand.
I stand corrected.
(sidenote, Google's Firebase acquisition has to be one of the best-executed devtool acquisitions right? Nobody would've been surprised if they had just gobbled Firebase up, spread the team across GCP, put the product on life support and after a year told customers to just move to GCP proper, somehow. Instead they realized that "GCP Easy Mode" was worth having a separate brand for and they executed excellently on that)
It’s insanely lean to get started. But at the end of the day, its GCP underneath with GCP billing (i.e. no spend caps). The easy mode UX, and the risk of real expenditure, sit in contradiction to each other. Of course, there a very generous free tier to placate that. But things like realtime need to be used with real cost considerations, and not frivolously once you need a feature only available on pay as you go.
Either way, they are all rather broad products and many users only use subsets. Thanks for speaking up, appreciated.
If i'll do something, will let you know if you wish
It's mostly about communicating the intend and a testament to Firebase establishing this rather broad product category, i.e. DB + dashboard + auth + files + ... You could probably also call it "a bucket of generally useful stuff", just doesn't have the same ring :).
> create per-user (likely with some limit of total 10 or so instance) demo env - average sysadmin like me should be able to handle it.
Totally feasible. There are also kubernetes deployment manifests in the repo. That said, I don't mind the shared instance too much. There's pros and cons. Happy to chat more.
> If i'll do something, will let you know if you wish
Sure thing. Thanks!
Does this project have a different focus than Supabase regarding the features to be implemented?
Supabase: https://supabase.com/
A Java/Kotlin client is certainly on the short-list. We recently added one for Swift. There are also clients for Dart, TypeScript, and C# in case you're more into cross-platform app development (Flutter, React Native, and Maui/Avalonia respectively)