What comes out
One ZIP named after the channel: a transcripts/ folder with a file per video named date - title, plus manifest.csv and manifest.json recording each video's status, caption track, and a SHA-256 checksum. A channel job's files look like this:
Practical Skills/ ├─ transcripts/ │ ├─ 2026-05-02 - How to learn anything faster.txt │ ├─ 2026-04-18 - The long-form interview.txt │ └─ … ├─ manifest.csv └─ manifest.json
How it works
The tool lists the channel's public uploads (Shorts and live replays included), pulls each video's caption track — creator-written when it exists, auto-generated otherwise — and formats everything consistently. Results stream onto a live, bookmarkable job page: previously requested videos appear instantly from the permanent cache, new ones tick in as they're fetched.
Honest limitations
500 videos per job, newest first, with a one-click "next 500" continuation for bigger channels; 2,000 new videos per person per day (cached videos are unmetered). A cold 500-video channel takes 15–45 minutes because fetches are paced. Private, members-only, and captionless videos can't be fetched — they're listed in the manifest with the reason instead of disappearing.
Common questions
Does it include Shorts?
Yes — a channel's uploads include Shorts and live replays, interleaved newest-first. To fetch only long-form uploads, take the channel's uploads playlist ID (UU…) and paste it with the prefix changed to UULF; the Shorts-only version is UUSH.
Which formats can I pick?
TXT (clean, timestamped, or both-in-one), SRT and VTT subtitles, Markdown for Obsidian and Notion, or timed JSON for pipelines.
Do I need an account?
No account, no email, no payment. The FAQ lists every limit in plain numbers.
Related tools: playlist transcript downloader · Shorts transcripts · the full channel guide