hey hunters 👋 i'm Kaivan, and I built Trove because I was tired of two things: 1. yt-dlp's command line being faster than I can think of useful flags 2. every "video downloader" website wanting my email + a subscription + half my screen as ads Trove is a single self-hosted webapp. paste a URL, hit save, get the file. that's it. no accounts, no download limits, no telemetry, no anything. runs on your machine. under the hood: yt-dlp + Flask + htmx + Tailwind v4. 🆕 in v1.0 (today): • 3× faster YouTube downloads via concurrent HLS fragments • IDM-style pause/resume that survives tab-close + server restart • persistent queue — in-progress saves come back on restart 🛣️ next: • MCP support — let Claude / your agent fetch files for you • smart categories (auto-tag by source / type) • one-click browser extension to grab absolutely anything from the web + livestreams. 1000+ sites supported (anything yt-dlp can do). MIT. ~3MB Docker image. setup is one command. → repo: https://github.com/afk1997/trove would love your feedback, especially what feels missing for your "save things from the web" workflow. reply with anything! 🥁 it's launch day, every upvote, helps a lot 🙏

Trove - Download videos from 1000+ sites

Paste a link. Save media locally. No ads, logins or limits

Paste a link. Save media locally. No ads, logins or limits