New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
A little video preview of the AI terminal, can be used for way more complex stuffs ofcourse.
For me at least it's quite useful to manage large fleets, I create custom detailed prompts and then it can run it across X amount of servers for monitoring/reporting to me specific things without having to set up a monitoring bloatware into the servers.
Trust is not just about determining whether there's malicious code. Vibe-coded slop is infamous for its severe security vulnerabilities. Despite the fact that LLMs are good at finding bugs that humans write, it's terrible with discovering the bugs it creates itself.
Why are you showing it off like a keylogger
Wym?
That's true, as I said you need human thinking and a good architecture made by humans to create quality code. People like you that thinks that in few prompts you can create a terminal like this are just delusional.
I mean if you have a bit of trained eye you can understand immediately if something is slopped or actually thoughtfully designed.
For an actual SSH client, you need plenty of review. One person and Claude is not enough to create a secure SSH client.
Anyone can vibe-code an app that they feel is "above" the "other slop down there". Just because it's not bottom-of-the-barrel "gimme a terminal pls" in one prompt does not mean it's not slop.
Yes but i think comments about AI soapiness and potential security risk are very relevant. I first checked github was still alive and noticed you have a lot of issues reported but only 4 pull requests and last WindTerm version was released in March 2025: this is the reason why I am not using it although it looks and feel very nice.
I did not then bother translating those many issues created in asian languages.
I just did it now (thank you Google Translate!) and apparently many think WindTerm is dead and are promoting they own app
We just released it, ofcourse we don't stop the development here, it's gonna be a fully supported project and hopefully someone would want to contribute on the repo development as well
So why advertise it as if it's a fairly-polished, secure application for people to try out and use rather than an early pre-alpha that is bound to have a number of fatal security bugs and is in need of careful testing in a safe environment?
No one said it's a perfect product man...
Still waiting you to find a fatal security bug that Im not aware of
Think of it like coming up with a new medication in a home chemistry lab. Sure, it's cool. Sure, someone should do some analysis. Maybe it'll be the next penicillin, who knows? But you shouldn't go advertising it on TV just yet.
People aren't harping on this due to a lack of perfection, that's simply not it. The issue isn't that it might not be perfect (nothing is), but that I see absolutely no disclaimers or warnings, when traditionally a new security product would have big, fat, bold warnings. Instead, it's written like it's intended to get people to find it useful so they use it (an issue for a project intended to do something securely), with no mention whatsoever being made of the potential issues or lack of an audit.
Aaaand now it fits in the bag with all the other vibe-coded security projects. I was waiting for that to come out.
I'll bail for now and leave behind some advice: Code does not become less secure after a security audit finds bugs. They were always there. Think about that for a moment, and think about on whom the onus of proof must be placed.
How can you even get so mad at an open source project but still be happy about closed source products that companies sell you for a fat 10$/month of which you have ZERO knowledge of how those are built.
It's quite curious for me.
Leave alone the fact that you are just assuming stuff without giving one fact, which is quite rude.
The answer to that curious question is simple: I'm not happy with closed source products. In any domain where security is important, I prefer to stick with solid, audited, and open source codebases. For example, OpenSSH and tmux.
If you feel the need to point out that OpenSSH and tmux have suffered vulnerabilities as if that indicates hypocrisy, do note that I did not say formally verified with a mathematical proof of correctness, but rather solid, audited, and open source.
See what I wrote one post prior:
Perhaps you and Claude write secure code the first time, so secure that even questioning implied claims of security (in the same sense that offering someone a drink brings with it an implied claim of potability) would be offensive, but I doubt it. Now, if you are aware that there may be, and in fact likely are, serious security bugs, then ask yourself why you do not add a big fat disclaimer that the code is unaudited and has never seen even one hour of scrutiny from a professional.
To expand on the potable water analogy, imagine I told you that I invented a new water filtration technique and offered you and everyone here a drink, saying that it can purify sewer water and make it fit for consumption. Surely you would be quite taken aback if, upon you questioning me about how safe it was and expressing skepticism, I took offense and demanded that you show me the bacteria that were not filtered out, no? You might even begin questioning my judgement as a whole.
Take care.
Maybe you like this stuff better @forest https://github.com/ZacharyZcR/termius-exporter
If that's the type of software you like
Why is the UI a carbon copy?
The tabs are similar because that's genuinely a good UX, even tho that's pretty much the UI of all SSH Client not being built 15 years ago.
But the remaining 90% of the UI is just different.
Ah yes, the secure Secure Shell client, or SSSH for short.
an ssh client is just maybe the worst sw to ai-generate and offer to public imo. opinions will vary of course but not a thing i'd even consider (edit: using).
Outside all of the hate in this thread, it's actually not a bad project at all in terms of UI, code or anything else.
I stopped paying for Termius a little while ago, and I've been scraping along with Tabby and Bitwarden SSH Agent authentication - it's been fine but could be better because it's quite finicky to setup on new machines or maintain over time.
However, I don't think I'd personally trust this in production because the security model on the Cloud Sync isn't what I think most people would look for. From what I understand, local vaults are only protected by the OS Keychain and Cloud backups are done to a Cloudblast account with a server-issued key, meaning Cloudblast staff with access to the accounts server could decrypt all of my secrets and keys. If there was an ecosystem to selfhost a vault, or the security model changed, I'd 100% consider it. Bonus points if there becomes support for extending it with plugins so we can add things to it that might help one specific workflow that wouldn't necessarily benefit the app to have as a feature.
Correct me if I'm wrong on the above, just had a quick look.
@cloudblast thanks for sharing. This place can be so harsh sometimes. I can't say i have a huge usecase for it at the moment, but it looks like a neat product and I'm glad you decided to share!
The cloud sync you are correct, we just kept this way for now for simplicity and ease of use, we plan to add E2E soon as a layer on top.
For the local version you actually have the OS Keychain + if you set the app password from the settings it will use that password as a passphrase to encrypt all the hosts a second time, making it quite safe.
You just made literal copy of Termius.
I see the vibe code resistance is strong here hahahaha
Nothing wrong with it unless you think its 100% perfect. It's not.
What it is is really useful tool for those that already code
Good luck with the project ill take a look.
Reminds me same kind of discussions about WordPress: developers arguing better to write everything from scratch because optimized, faster, etc. vs people who don't know how to code and only want to quickly build a (nice) website
Of course security is no joke and should be taken very seriously
Good idea, bad implementation. If it's javascript/electron, just make it a web application that you can self-host. I hate web applications pretending to be a universal desktop app.
If the display is rendered by a web container like Chromium but runs locally, it's not a desktop app but a crippled web app.
Looking forward to it. Not a bad product at all. My opinion is nothing is wrong with AI assisted development (informally "vibecoded") as long as it's limits are acknowledged and it's monitored to make sure the output is high quality, same as any other code-peer-review process in any other team anywhere.
Product itself looks really nice and is a really nice looking version of what I'm to use, so I'll give it a spin once work starts on proper E2EE or I can selfhost the server.
Good luck with the project!
Thanks for the considerations, this could actually be an idea for anyone that wants to build the servers version of the vault, I can then implement the custom endpoints entry form in the settings to choose something else over the cloudblast auth.
I'm also open for any other provider who wants to join and offer the vault hosting to be added as an option.
no body need more ssh client.
is windterm vibe-coded?
website is: https://windterm.io/
Latest commit on the source is 5 years ago, so I believe is not