update 2026-03-12 - I discovered the command line switches to say "best that isn't webp"

Offline archiving of YouTube videos is possible with this "feature-rich command line tool" called yt-dlp. Maybe you suffered a local data loss and you want to recover your uploads

Or maybe you just don't trust companies anymore and want to have a local copy of everything you love before it's all taken away from you. You do you 👍🏻, I won't judge.

Here's a command I've cooked up that helped me archive a channels worth of content in a date-sortable list.

$ yt-dlp -f 'bestvideo[ext!=webm]+bestaudio[ext!=webm]' -N 4 --cookies-from-browser firefox --output "%(uploader)s - %(upload_date>%Y-%m-%d)s - %(title)s [%(id)s].%(ext)s" https://www.youtube.com/@TimLNX/videos
Extracting cookies from firefox
Extracted 574 cookies from firefox
[youtube:tab] Extracting URL: https://www.youtube.com/@TimLNX/videos

Why those option flags though?

  • .webm doesn't integrate well with my current tech stack

Using -f 'bestvideo[ext!=webm]+bestaudio[ext!=webm] will select the best format that isn't 'webp'. I have no personal issues with that AV container, it just doesn't integrate well into my stack. MP4s (or whatever else I get) are seamless, so I am archiving in that format.

  • Archive faster, maybe

Attempt concurrent downloading with -N 4 to download 4 video fragments simultaneously. Anecdotally I think this has got me throttled before, but that could be a false memory. tl;dr - try turning this option off, or changing the number to see what you can get away with.

  • Save the archive sortably

I would like my archive stored in a way that is easy to sort chronologically (by date uploaded). The default yt-dlp behavior will save videos with file names like <video title> [<video id>].<ext>, and that doesn't sort very well. That is why I use this format instead: --output "%(uploader)s - %(upload_date>%Y-%m-%d)s - %(title)s [%(id)s].%(ext)s". Thanks to this post for saving me the time to learn the output template variables myself.

  • Authenticate as myself

By passing in --cookies-from-browser firefox I can interact with the YouTube API as myself and save my private unlisted videos. If you see a message like ERROR: unable to download video data: HTTP Error 403: Forbidden then that's probably what you're running into.

Sometimes stuff just won't work. On Mac I find a brew upgrade usually fixes the issue.