Original Summary

VidArch started as something I built purely for myself. I wanted to keep some videos locally, but I never really liked the usual workflow: find a video somewhere, copy the link, open a downloader, paste it, start the download, then end up with a file that has basically lost all the context around where I found it. I wanted something closer to browsing a normal video platform, but built around discovery + local archiving . With VidArch, I can search for videos and channels directly, browse around, subscribe to creators, and download whatever I want to keep without leaving the app. Once a video is downloaded, it becomes part of the local library while keeping its thumbnail, channel, metadata, history, playlists, etc. Over time I added the stuff I actually wanted to use myself: local search + remote results channels and subscriptions downloads with quality selection, queue and progress auto-download for selected channels local library, playlists and history importing videos that are already on the server built-in player with seeking, playback speed, PiP, etc. Docker Compose deployment One thing I cared about quite a bit was separating two different needs : discovering/archiving content and using your local library . VidArch has a local mode where the remote features can be disabled completely while the library, archived channels, playlists, history and playback still keep working normally. I didn’t want the local media side to become useless just because the discovery side is unavailable or disabled. Also, just to be transparent : I used AI quite a lot while building this . But it wasn’t a one-shot “vibe code it and ship it” project. I spent a lot of time thinking through the product, changing the architecture, reworking features, testing workflows, and dealing with things like persistence, Docker, permissions, security, updates, migrations, etc. AI was a development tool, but there were a lot of iterations and decisions behind the final result. If there are developers here who feel like looking through the code, I’d genuinely be interested in constructive feedback too, especially around architecture, security, or things you would have done differently. A quick note about the demo: it’s only meant to give you a feel for the interface . Remote search and the yt-dlp/download side are disabled there, so it’s not the full product running publicly. Those parts only work on a real self-hosted instance. The project originally existed only for me, but at some point it became complete enough that I wanted to clean it up properly and make it public. Longer term I’d like to add proper multi-user support, with separate histories, playlists and preferences. I’m also considering making a mobile app and maybe a desktop app later if there’s actually interest in that. I’m mainly curious what people here think about the discovery → download → local library approach, and what would still be missing for you to actually use something like this. GitHub: https://github.co

  • 情报分类：商业与市场研究
  • 分类依据：内容涉及商业、投资或市场动态
  • 信息来源：Reddit · SideProject
  • 发布时间：2026/9/24 18:35:37