briffy

The project I posted here yesterday focuses on providing text, voice and screen share. My goal is to provide an easy to host tool for those three things. Check it out if it's just those you want in a single package.

I think we can all agree coming up with names is not one of my key skills. Not following standards, however, is. Hold my beer, I got this.

It's a new project that I've been working on locally. My code backups are handled locally (this is /c/selfhosted right?) so I've not had a reason to commit anything to GitHub until I was ready to let other people have a crack at it.

My Github account is 14 years old and has previous projects on it with 40+ stars. I have been actively replying as much as I can in this thread and the project is entirely open source, feel free to review the code and let me know if you find anything suspicious in there.

Thanks for sharing, that's an interesting read. I hadn't come across this when I was looking (to be fair, it might have been posted after I looked, I've been mostly on stack overflow for a month).

I'm in agreement with many of the author's points (I've ticked all of their requirements except markdown support and the only stretch requirement I don't meet is being able to scale up to thousands of users - but I never wanted to do that anyway).

I am really torn on "everything needs to be federated" though. I feel like credential fatigue/ease of joining a server is largely solved with SSO/SAML or magic links with guest access. I want to love federation, I really do... But my own lived in experience with matrix has soured me on it. It was a pain to maintain and the eventual tipping point was one of my "trusted" federated services (Arch btw) flooding me with CSAM.

I think there are many discord users that use it for voice/text/screen share with a core group and don't really care about all the extra stuff or having these huge servers filled with people they'll never interact with. It's just the only realistic option right now to chat with their friends. Those are the people I'm hoping to attract. I'm not saying what I'm offering is a perfect solution but what I am offering is that core functionality without the gradual enshittification and constant slurping of your data.

I'm with you, I love Matrix as a concept but the experience of actually running it was a major headache for me.

I'd love to contribute to those projects but anyone that's read through my repo for this will see I'm not that good. It took me a long ass time to figure out end to end encryption and those projects are built on it. 😅

I also feel like they fit a different niche, at least matrix does, I'm not too familiar with XMPP. I've said in other replies, I'm not looking to make something that's infinitely scalable or federates with other services, just a relatively simple chat app that someone can have running for their group of gamer friends. If it can do text/voice/screen share with minimal setup/fuss/external dependencies then I'm a happy boy. I kinda had this idea in my head that I'd like to get it to the point where you can upload a tar.gz to cheapo web hosting, untar, follow the setup wizard and have comms ready to go without having to mess around in config files and what not.

This will probably be the first update I release. I've pre emptively built the front end in Quasar and there's even some bits of commented out code in there from where I started looking at storing authentication data per server. The plan as I see it, and I think this makes sense, is to have the web app based front-end be for people that want to run their own contained instance of both the API and front-end but then also have a Quasar desktop based app that has server switching built in. This then allows the server owner to just run the API if they want and let the user worry about how they connect.

It hasn't been a priority for me at the moment because I'm literally the only person running a server. 😅

Now that it's out in the wild, my next focus will be on the multi server side of things and making the text channels a bit more functional than just plaintext.

So the group channels and audio/voice aren't but DMs are. It uses asymmetric signing and per conversation keys. These can be imported/exported so you can see your conversations across devices but by default the keys are never transmitted.

Unless there's an issue with my code I'm missing?

Edit: oh wait, this was a reply about fluxer, I'll leave it up just for info in case anyone is interested. Can you tell this is day one using Lemmy... 🫠

Defending my work in a public setting is not being unable to emotionally handle criticism. Framing it that way is disingenuous but I think you know that and just want to push your anti AI agenda.

Sorry, you can call me emotionally unstable all you want but if you think generating the template for the GitHub readme (not even the install instructions or anything, just the template) and some favicons invalidates hundreds of hours of work then it's you that needs to do some reflection.

There is a reason I am concerned about "being right". This is a project I've worked pretty much non stop on for about a month, have written every piece of code myself. This is a public forum and the first time I've put my repo out there. To have the very first response be a dismissal that the literal hundreds of hours I've put it into it is just AI is not only insulting, it also makes it difficult for me to get valid feedback as people won't read past that first comment and actually look at what I've made.

Sorry but I won't roll over and take it when my hard work is dismissed because I used AI to generate the GitHub readme template. That is absurd.

Thank you for your feedback.

Honestly, there are a few more features I want to add like markdown/gif support, image/video uploads and user settings (push to talk, volume leveling, etc. are sorely needed) but outside of that and cleaning up bits of the UI, I'm terrified of creating an unwieldy codebase. I don't want something that requires constant tweaks and updates, just something that focuses on those core features and nails them down. You can just install it and pretty much forget about it.

Docker is definitely on my list, I was going to have a chat with one of my mates that lives in docker land to see if he could give me a hand setting it up properly. I've used docker before but mostly in hacky ways rather than something I'd be comfortable deploying to other people.

Appreciate the feedback, but if I'm going to be quoted, also me:

The actual meat and bones of the project, the code, is all written by me, a real human.

Let's call a spade a spade here. It was not a legitimate GenAI concern. I considered rebasing the repo after I realised my mistake but I honestly thought no one would care, or the people that would care enough to look at commit histories would know enough about what they are looking at to realise what I did.

What OP did was look at the commit history, not understand it and then accuse me of being AI. So no, that is not in good faith.

Thank you so much. I've run through the setup so many times and got a few instances of it running but I also understand how it's all working under the hood so I have a home field advantage. It would be a massive help to have someone go through the setup and make sure it's actually doable by someone that didn't make the thing because I'm so paranoid I've missed something in the instructions and it just simply doesn't work.

The poll rate on the audio is pretty aggressive, will be interesting to see what happens if your requests take longer than the poll rate of sending/receiving audio. I've accounted for packets arriving out of order so should be okay but... We'll see. 👀