The workflow: manually re-watching videos to pull quotes. The fix: one command.

The broken workflow: someone on the team watches a competitor demo, a recorded customer call, or a conference talk, and manually scrubs the timeline to find the three quotes or feature mentions worth pulling into a doc. Repeat every time a new video shows up. Nobody owns it, everybody does it badly. The fix is boring on purpose: yt-dlp, a free command-line tool, pulls audio, video and subtitles off more than a thousand sites (YouTube included) and writes structured metadata, title, description, chapters, alongside the file in one command. Point it at a link, get back an audio file that's ready for a transcription model and a JSON file with everything you'd otherwise copy out by hand. It's not an agent, doesn't try to be one. It's the plumbing step that makes the actual agent work (summarize this call, pull the objections, draft a follow-up) possible in the first place. The workflow this replaces was never the transcription. It was the twenty minutes of manual scrubbing before the transcription could even start. Repo: https://github.com/yt-dlp/yt-dlp Bring your version: what's the workflow on your team that's still "someone watches the whole video and takes notes"? I'll suggest where a tool like this slots in.