[QA] - project status

[ciltocruz]

Simple: Is this project abandoned?

[gazpachoking]

We have minimal maintainers on the core, but there isn't anyone left who was involved with the webui, so it's in a bit more precarious situation.

[gazpachoking]

I've been considering starting a v3 if the webui which is simpler and more maintainable in pure python. I'd love to ditch react, npm, build steps, the json API, and most JavaScript. I already don't have much time for the project anymore though, so I'm not sure how much better of a situation that puts us in.

[logicalChimp]

Hmmm - there's a lot of existing content that would have to be recreated if you build a V3 webui...

It's been easy enough (so far) to use e.g. Claude to do some minimal development (to add a couple of plugins to make it easier to grab the series / episodes I want)... so that might be a viable option too.

(speaking as a python developer myself - I can broadly read Typescript, but I can't easily write it and have no idea about the libraries... getting an AI to do the scutwork seems a viable option, if there are no actual web-devs available).

[gazpachoking]

Yeah, I still need to look at what we have in the existing webui still and evaluate how much work it would be, I'm not even sure how much is there. Part of my motivation is it would actually make the work a bit interesting, whereas I have no desire to interact with the current codebase at all. I am not even running the webui at the moment, but after the enshittification of trakt I'm looking for a replacement to control my FlexGet instance now.

[ciltocruz]

The new idea sounds good, but I think it's very difficult to implement.
I opened the UI today after two years and saw that it was still the same. It's complicated to manage in terms of the database, slow, and visually outdated... that's why I asked.

[gazpachoking]

Hopefully I'll do a quick POC coming up to see how well it would work and for much work it would be.

[paranoidi]

Considering how well agentic implementation works, I think it would be entirely feasible to drive this purely from prompts. I have thought about trying that for months but I have been spending my tokens on elsewhere meanwhile.

Although I do love the simplicity, I don't think purely backend generated UI feels very nice these days and implementing more complex editors for example would be much harder than just integrating some react component.

[gazpachoking]

Possibly. Have you looked at Datastar? I have been using it for some work projects and I think it enables all the nice interactivity typically associated with SPA frameworks but feels so much simpler and nicer to write. You could be right about more pre built things being available for react though.