Collection Sharing (Discord Server)

Currently, building a collection is fine and all and in theory it can be shared, but there's nowhere to really share them. While it'd be most ideal within the app itself, that comes with a whole host of issues, such as hosting a server, having the app actually have internet access, etc. As a measure in the meanwhile, perhaps a discord could be setup where people can discuss, share collections, etc. Given that this version is closed source, that would be the most ideal place for everything for the time being.

Text Detection

Given that there already is a feature to have something always visible regardless of the censoring around it, I feel like it should be expanded to detect text as well. Currently, captions, videos with text, just anything with text will get covered by the censor, which isn't a huge deal, especially if you adjust your censored areas to be smaller, but the default 100-110% for example usually doesn't catch stuff well enough, so it's a necessity in my experience to have them sit at 150-200%.

Multi-monitor support

Specifically, multi-monitor support with only one instance of the application running, I can launch multiple instances and cover multiple displays, but that's quirky, leads to having to setup the entire collection beforehand and tweaking it means closing the others, only keeping one open.

Using your own model

Not sure how feasible it is, no clue what DirectML expects, but I was really hoping that I could just drop in a different .onnx model for detection and it just crashed. I'm not bothered by using the shipped model, but it'd be a nice option to have.

Visuals

Not really something that I think needs to be a priority just yet, but a major gripe for me is the use of icons for "Triggered by" and a similar selection and toggle outline. More contrast basically between the elements, color or otherwise, would be great.

Hey. Thank you for these very nice ideas !

Collection sharing : I agree that a Discord server would really benefit to Hotscreen because people could share lots of very cool filters. But I don't want to manage it and I fear that people could share things that would ban Hotscreen. I could create a Discord and make some poeple moderators : maybe you or theguywhodidit ?

Text detection : I could add that to the detection model now that I can train my own model. To avoid censor being above text if I understand well.

Multi-monitor option : I will have to do it but this is a lot of work since some screens can use different recording methods so every parameters must be rethink.

Using your own onnx model : For now, Hotscreen expect a very precise model (input and output). After the next update, you will be able to select between models that I created. Maybe I will do a tutorial to finetune it with your own images.

 Visuals : You would like to better see which bodyparts are selected and not all in grey icons ? I'm very bad with visuals I made those with Stable diffusion. I will try to improve that.