KeePassχ - a KeePassXC fork
18 points by jmelesky
18 points by jmelesky
Previously, on lobste.rs. I'm not sure if I'm a fan of the "no AI fork, created 6 hours ago" genre of post here, but once again I think whitequark is competent enough to work on something like this.
I wonder if they'll fix the UI issues. I have switched to GNOME Secrets because of KeePassXC's AI usage, but it also shares a similar issue (and several worse issues). At this point I think a CLI password manager would be a major improvement in usability over the GUI ones I've used in recent memory, which is a funny thought.
Can GNOME secrets do passkeys? Im having to use Bitwarden for just one site that needs a passkey. I couldn’t find a native linux/firefox solution for passkeys
I'm as skeptical of LLMs as the authors, and prefer smaller human made software.
That said, the confidence I have built in the KeePassXC project by using it for many years massively outweights their stance on LLM usage.
As of today, it has always worked really well and felt solid.
Good initiative, but why was this posted now, without any meaningful changes vs. the release from a year ago? Are you looking for contributors?
I don't know what to think of this. The KPXC devs had to mass-delete abusive stuff in their dev channels after the first AI-Hitlist, in a sense this is another hit piece. The KPXC-Devs clearly don't just blindly use AI while having shown at previous audits that they are capable of properly doing the encryption parts (the audited part).
It's easy to claim that KPXC is just supposed to be stable and just port it to QT6 when in reality people have been asking for features like Sharing, Yubikey, Passkey, SSH & OS-Wallet integration - and all of those were accepted as defacto standard for modern password managers. There are far far simpler password managers if you really want that - and there is the original Keepass2.
It's hardly a hit piece. All issues about code quality aside, there are still huge ethical reasons to take an anti-AI position. The fork will succeed or fail in the traditional open source manner, I think it was inevitable, and I think having the choice is great for the wider community.
I saw a proposal for an SQLite based KeePass* family database format, since the existing XML format is such a pain to work with.
It was on here, I wish I could find it again because I do think it has merit and it’d be nice to see some folks experimenting with it.
