YouTube Transcript Scraper: Channels, Captions & Subtitles
Pricing
from $2.40 / 1,000 transcripts
Get the transcript of any YouTube video, Short, channel, playlist or search results as text, timestamped segments, SRT, VTT or Markdown, in any available language, with title, channel, duration and views. Videos without captions can be transcribed with AI (Whisper). Pay per transcript.
Video links (youtube.com/watch?v=..., youtu.be/..., Shorts, live, embed) or 11-character video IDs. One per line. Channel and playlist links pasted here also work.
Get the newest videos of whole channels or playlists: @handle, youtube.com/@handle, youtube.com/channel/UC..., or a playlist link. Combine with Only new videos and a schedule to get every new upload as text.
[ "@TED", "https://www.youtube.com/@NASA"]How many of the newest videos to take from each channel or playlist. Listing is free; you pay only for transcripts returned.
Get transcripts of the videos YouTube shows for a search, e.g. "python tutorial" or "AI agents". Results keep YouTube's order (searchRank). Combine with Only new videos and a schedule to get new videos about a topic as text.
[ "python tutorial"]How many results to take from each search (20 per page). Searching is free; you pay only for transcripts returned.
Order of the search results.
Only videos uploaded in this period.
Only videos of this length.
Skip videos of these channels, playlists and searches that earlier runs already processed. Ideal on a schedule: each run returns (and charges) only new videos.
Language codes in order of preference (en, es, pt, de, fr...). 'es' also matches regional tracks such as 'es-419'. Empty = the video's own captions.
If none of the preferred languages exists, return the video's captions in its own language (billed). Turn off to skip those videos instead (not billed).
Use captions written by the creator when they exist, and auto-generated ones otherwise. Turn off to prefer auto-generated captions.
When a video has no captions at all, download its audio and transcribe it with open-source Whisper (99 languages, auto-detected). Billed $0.010 per minute of audio instead of the transcript price; each video is rounded up to the next minute. Off by default. The run uses 2 GB of memory when this is on.
Videos without captions that are longer than this are not transcribed with AI (and not billed). Protects your budget from very long videos.
text (paragraphs), segments (start/end/text), srt, vtt, markdown (with [HH:MM:SS] timestamps) and chunks (for RAG).
[ "text", "segments"]Channel ID, duration, views, keywords and description (free).
Target size of each chunk when 'chunks' is selected.
How many videos are processed at the same time.
Mark the run as FAILED when a video fails after all retries or no transcript is returned. Videos without captions do not count. Useful for monitoring pipelines.