1. Album collaboration: https://ente.io/blog/collaborative-albums/
2. Photo collection: https://ente.io/blog/collect-photos/
Please let me know if I missed something. Sharing is an important feature for all of us, and we would love to get it right.
2,085 karma · joined November 23, 2016
1. Album collaboration: https://ente.io/blog/collaborative-albums/
2. Photo collection: https://ente.io/blog/collect-photos/
Please let me know if I missed something. Sharing is an important feature for all of us, and we would love to get it right.
We currently keep 2 versions of a photo - one the original, and the other a downscaled copy to be rendered as the thumbnail.
Unlike non-e2ee providers, we cannot transcode and serve optimised images on the fly, when it's faster to downscale than serve the original image over network.
What we could do is
1. Intelligently preload original photos when their thumbnails are in scope
2. Store an extra version of the photo, whose resolution is between that of the thumbnail and the original, and perform #1 over those
Sorry about the flicker, will fix it.
Now Ente could of course choose to keep lesser replicas and offer "lite" plans that are more affordable. But we would rather not complicate our pricing structure right now. Understanding buckets of GBs is hard enough, and adding tiers on top would worsen the experience for most.
All of that said, Filen does seem like a really cool project for storing files.
We do get your perspective. Privacy does not outweigh convenience for vast majority. We are hoping to find the balance.
We have to pick our battles, and we currently aren't looking at serving customers on low-end devices. But if they access Ente on a laptop, the indexing will run there, and the computed indexes will sync to their low-end devices, e2ee.
Also, given that compute on smart phones is getting better and cheaper with every iteration, we believe it's best to bet on Edge ML for the long term.
The latter is very likely in the long run.
Most of our costs come from keeping 3 replicas of your data: https://ente.io/reliability
We could perhaps reduce the replicas from 3 (2 hot, 1 cold) to 2 (1 hot, 1 cold), and lower our costs, but it's not something we've actively thought about.
Also, our prices are designed such that the business can run sustainably, and we believe that's the best way to build Ente, where the expectation is for the company to outlive its customers.
Some bits of the blog post are outdated, since the feature has matured since then. But the implementation details are roughly the same.
TL;DR: The keys are added to the URL fragment (the part that follows the #), and these fragments are not accessible to the server.
Google Photos already has a public API. But you cannot consume it if you are building a competing service[1]. Also, the APIs will not serve you files in their original quality.
If they wanted to honor the intent behind GDPR's portability mandate, it would have been trivial for them to enable seamless migrations.
Now what's ironic is how Google is a "partner" in the Data Transfer Initiative[2].
[1]: https://developers.google.com/photos/library/guides/acceptab...
[2]: https://dtinit.org
If there are any experts reading this, we'd be grateful if you could let us know what we're doing wrong here.
We're in the middle of switching over to SES because they are landing flawlessly into Gmail inboxes, but it would still be good to understand what we could have done better.
But it splits your metadata (like capture time, geo coordinates) and places them in a separate sidecar file.
This isn't an issue if you're migrating to Ente because during import the desktop app[1] will merge the metadata with their respective files.
If it's for you to keep a local copy, you will unfortunately need to write some scripts. There's in fact a paid product[2] that does just this.
Ente now has everything except cross-platform face-recognition (it's desktop only right now), labels and EXIF edits. The first two are being worked on, we should have v1s ready by Q3. Polishing will take a bit longer.
The long-press-to-skim-through albums is very neat, will add this.
I've scribbled a bit about the story behind the name and the mascot here: https://ente.io/blog/ducky/
Ente's strength lies in end-to-end encryption[2] and its cloud[3] offering so you don't have to worry about reliability.
So if self-hosting is what you're after, Immich, Photoprism and Damselfly (TIL!) are perhaps better designed to serve your needs.
[1]: https://github.com/photoprism/photoprism
You can read about our replication strategy here: https://ente.io/reliability
We could in the future offer cheaper plans at the cost of additional replicas.
We had opensourced our server[1] yesterday, which is perhaps why we are on the front page. Stoked to be here
Ente had launched on HN[2] a while ago and has been sustainably growing since.
We took the feedback from our Show HN seriously and have since
- undergone a cryptography audit [3]
- published our replication strategy [4]
- added requested features (family plans, collaborative albums and links, ...) [5][6][7]
- made progress with Edge ML [8][9]
- built a CLI for incremental data exports (our desktop app supports this as well) [10]
- and in general matured as a company [11]
Also, apart from our source code, our Figma[12] is public as well.
If you've feedback on what we could do better, please do share, it'd be very helpful.
And if you've any questions, do ask, I'd love to make myself useful.
[1]: https://ente.io/blog/open-sourcing-our-server/
[2]: https://news.ycombinator.com/item?id=28347439
[3]: https://ente.io/blog/cryptography-audit/
[4]: https://ente.io/reliability
[5]: https://ente.io/blog/family-plans
[6]: https://ente.io/blog/collaborative-albums/
[7]: https://ente.io/blog/collect-photos/
[8]: https://ente.io/blog/image-search-with-clip-ggml/
[9]: https://ente.io/blog/desktop-ml-beta/
[10]: https://github.com/ente-io/cli
[11]: https://ente.io/blog/reflections-on-trusting-trust/
[12]: https://www.figma.com/file/SYtMyLBs5SAOkTbfMMzhqt/ente-Visua...
The general inconvenience of waiting for emails aside, this flow also had the potential to lock out customers who were using Ente Auth to store their email's 2FA credentials. So it was important that we fixed this.
Once OPAQUE becomes more mainstream, we will very likely adopt it.
If you've any questions, please do ask, I'd love to make myself useful.
TL;DR: Removing Protobuf and directly serializing the list to a blob helped!
We'll take another look.