Show HN: Stop paying for Dropbox/Google Drive, use your own S3 bucket instead (locker.dev)

by Zm44 208 comments 252 points
Read article View on HN

208 comments

[−] dewey 38d ago
The selling point of Dropbox/Google Drive isn't the storage itself, but that there's app for mobile and desktop operating systems which deeply integrates it in the OS so it's just like a local folder that's magically synced.

So it's a cool project, but not really what I'd say is a Dropbox replacement.

[−] Tepix 38d ago
On the other hand when a Dropbox user shares a file with you these days, the nudges have so gotten out of hand that it's a pain to use.
[−] sumeno 38d ago
Why wouldn't I trust a vibe coded app that has existed for 1 week with all my important data?
[−] charles_f 38d ago
Stop paying for banks, AI built this cardboard box that you can store in your toolshed instead!
[−] filleokus 38d ago
Neat! Pricing wise it might not always make sense though to use the commercial blob storages, especially for solo usage.

1 TB is roughly 20-30 USD per month at AWS/GCP only in storage, plus traffic and operations. R2 is slightly cheaper and includes traffic.

Compared to e.g a Google AI plan where you get 5 TB storage for the same price (25 USD/month) + Gemini Pro thrown in.

[−] kristianc 38d ago
We've officially come full circle

https://news.ycombinator.com/item?id=9224

[−] ovaistariq 38d ago
The critical part of Dropbox is not just the storage layer but a combination of their client and server. Even small things like how do you handle conflicting writes to the same file from multiple threads, matter a great deal for data consistency and durability.
[−] 1a527dd5 38d ago
Absolutely not. The value isn't in the cloud storage. The value is in the client (DropBox in my case) seamlessly working across all my devices.
[−] rkagerer 38d ago
Why would I want to replace my reliance on them with reliance on Amazon or another cloud provider?

I'd rather control the whole stack, even if it means deploying my own hardware to one or more redundant, off-site locations.

Edit: Are there robust, open source, self-hosted, S3-compliant engines out there reliable and performant enough to be the backend for this?

[−] xnx 38d ago
Features I'm guessing this is missing (in no particular order): Recycle bin, Sharing with permissions, Editor, Versioning, Search, Partial sync, zone redundancy/backup, Windows, Android, Mac, and iOS clients
[−] ks2048 38d ago
I pay Dropbox $120 per year for 2TB. No transfer fees, solid Apps, macOS integration, free APIs.

How much on S3? A LOT more.

[−] cjonas 38d ago
Why does getting started have me sign up for an account vs take me to the docs to self host?
[−] gargan 38d ago
May I recommend the excellent https://s3drive.app/ which is compatible with S3 and also providers like Proton Drive
[−] npodbielski 38d ago
I bought 35$/mo 16TB server from OVH. I am running 2 replicas of Garage, one on this server. I am using this for backup for now but probably I will also move my Nextcloud files there and websites. This is fine for now and less pricey than any S3 provider I was able to find.
[−] SergeAx 38d ago
If you are not ready to trust a vibe-coded app with all your digital life, I recommend Filestash[1], easy to install self-hosted frontend for virtually any type of storage. Written in Go, it can be enhanced with plugins. Uses local SQLite database.

I am using it with Hetzner Storage Box[2], which is insane value for money at 11 euro per 5 terabytes per month.

[1]: https://github.com/mickael-kerjean/filestash [2]: https://www.hetzner.com/storage/storage-box/

[−] pocksuppet 38d ago
Why would I use S3 when S3 is super expensive compared to all other cloud storages?
[−] iamcalledrob 38d ago
The actual best open source Dropbox replacement already exists: SeaFile. Too bad their website is terrible.

https://www.seafile.com/en/home/

It's pretty magical. It nails the "online" vs "cloud only" paradigm via the SeaDrive client. I have it running on my file server, and now all my machines have access to terabytes of storage with local performance, since it can cache a subset of your content locally.

And since I can run the server on my LAN, the throughput is way better than Dropbox would be too.

[−] vitalscope 38d ago
That is a bit like saying “Don’t use a medical analysis app, just interpret your lab results yourself.”

Sure, ChatGPT can help, but to use it reliably, you still need enough medical knowledge to ask good questions and evaluate the answers.

[−] vgr-land 38d ago
Looks like a good light weight solution to front object storage with a front end and auth. One suggestion is to add the license to the repo. The readme says License: MIT, but there’s no license file.
[−] charles_f 38d ago
I used to be excited by these kind of tools, I love to self-host stuff. When I clicked on the link, I had this hesitation, suspecting "maybe it's LLM generated". And sure enough, coming back to HN, description says it is.

File sync can't be that hard! Enters the first 3 way conflict and everything explodes.

Dont misunderstand me, this is a cool idea. But if your rotation time between ideating a project and pushing it to HN is a week, you don't understand the problem space. You didn't go through the pain of realizing its complexity. You didn't test things properly with your own data, lost a bunch of it and fixed the issues, or realized it was a bad idea and abandoned it. I have no guarantee you'll still be there in a month to patch any vulnerabilities.

Not that any open-source project had these kind of guarantees until now, but the effort invested in them to get to that point was at least a secondary indicator about who built it, their dedication, and their understanding of the space.

[−] carefree-bob 38d ago
You can do a lot with cloudflare tunnels and a home mac mini. Dropbox, blogs, homebox (home inventory), hosted git.

I'm a kid in a candy store playing around with this stuff.

[−] lain98 38d ago
I use archive storage class on google cloud, to store old movies and wedding videos, pictures of old vacations.

For everything else I use paid onedrive subscription. The biggest problem is user interface with s3 like storage and predictable pricing because remember you also pay for data retrieval and other storage apis, with dropbox etc you pay a fixed amount. Every year or so I roll over data into the bucket.

But for infrequently accessed data its fine.

[−] hodder 38d ago
The irony of this SHOW HN is that "just roll your own" was the classic OG critique of Dropbox before they took the space rocket to unicorn land.
[−] JackeJR 38d ago
https://news.ycombinator.com/item?id=47676174

Dropbox is a lot more than file storage. The syncing itself has been through serious tests to characterise its behaviour. Sure, some may not like the decisions taken to direct its sync behaviour one way or another but at least all these are known through formal testing.

[−] crazygringo 38d ago
"Stop paying for"?

But you have to pay for your own S3 bucket as well... and it's generally several times more expensive per terabyte, though this depends on different factors. (Not to mention you might still have to pay for Google if your e.g. Gmail doesn't fit into the free tier anymore.)

If this is supposed to be financially motivated, the creator seems to have it somewhat backwards.

[−] dboreham 38d ago
For all the people pooh-poohing this, I'm very interested in this business model (bring your own provider token) and this looks to be nicely done. I'm going to try it out. In particular I want to see if it supports payload encryption for the data in S3. I'd need that to be comfortable stashing all my personal data in AWS or Wasabi.
[−] robotnikman 38d ago
There was a similar program I used when I was working at a smaller company and built an automated backup system for their call center recordings. It basically mapped the S3 bucket to a Windows drive letter, since the PBX call recording software was running on a Windows server. It was a while ago so I can't recall the name of the program
[−] jrochkind1 38d ago
I wonder if it would be possible to do something like this that had transparent end-to-end encryption.
[−] ge96 38d ago
Funny I kept getting this "you are using 98% of your storage space" message with my 15GB so I'd be finding/deleting old attachments then eventually I was like fine I'll pay, it's like 48 cents a month or something
[−] artk42 38d ago

> Languages JavaScript 55.2%, TypeScript 44.4%, Other 0.4%

Isn't that compromises the whole purpose of the project immediatly?

Moreover, any reasonably adequate dev would work on expansion of syncthing ecosystem, not inventing a rasclet instead of a wheel

[−] pluc 38d ago
I'd love a local offline alternative, maybe I'll get AI to build it for me
[−] nunez 38d ago
The thing I find interesting about apps made by Claude et. al. is that they always fallback to using dotenv for configuration. I thought dotenv was on its way out! Personally, I've been using sOPs for this purpose.
[−] conradev 38d ago
I love it! How would you position yourself relative to existing OSS products in the space like Filestash or Seafile? I'm trying to pick a solution at the moment, and the mobile experience matters a lot to me.
[−] aitchnyu 38d ago
What happens if the server disappears permanently and only the bucket is up?
[−] robertclaus 38d ago
Feels like this is missing some of the key points of using generic bucket storage for me: 1. Archive pricing for really large old documents. 2. Cross-provider backups; especially for critical documents.
[−] itherseed 38d ago
Ok, I'll see it later but please use the 'Release' feature of GitHub. It is the easiest way to tell for your customers that a new release is out. Even GH can send notifications. Thanks.
[−] WhitneyLand 38d ago
Wouldn’t this be way more expensive?

Example 2TB:

Google $10/mo vs S3 ~&45/mo?

You could get cheaper that Google Drive with glacier tiers but that’s a different level of restrictions and still has retrieval fees.

[−] randomtoast 38d ago
The repository only exist for seven days and was likely written by Claude code, which makes it not very trustworthy for storing personal data.
[−] bovermyer 38d ago
Very cool idea, but without background file syncing from/to my local machine, it can't replace my cloud storage provider.
[−] angelfangs 37d ago
I use a cheap vps + dufs. Great for drag and dropping files and have my jellyfin server pointing to those directories.
[−] Isolated_Routes 37d ago
What is the privacy architecture? I like the concept, but am curious what safeguards you have in place.
[−] adamgordonbell 38d ago
Looks great!

Feature request: Google Drive for desktop.

That is the feature that gives your drive as a mounted file system that stream files as you need them.

It gives me the ease of having access to a giant amount of files stored in my gdrive without having to worry about the space they take up locally nor moving files up and down.

Actually, what solutions to that might already exist? I don't really use the web UI of gdrive as much as use it as a cloud disk drive.

[−] huijzer 38d ago
But isn’t Syncthing already open source Dropbox? Can easily use own hardware too which is very nice.
[−] tobi_bsf 38d ago
The magic of Dropbox lies in its local app and sync, this is a nicer webinterface for s3 storage?
[−] BorisMelnik 38d ago
I already do this with next cloud and s3, I've never once had an issue in many years
[−] novoreorx 37d ago
S3 and B2 both have web UI, so if this is what your vibe app do, it's useless
[−] noja 38d ago
Why not just use an FTP server?
[−] NoahZuniga 38d ago
Google drive pricing is cheaper than S3/Google cloud storage for storage.