YouTube Transcript Scraper — RAG-Ready Text & Subtitles
Pricing
from $5.00 / 1,000 transcripts
Used by RAG builders, content teams and researchers turning video into searchable text. YouTube transcript scraper with chapter-grouped, chunk-ready paragraphs for LLM and search pipelines. Re-runs skip videos already delivered so a corpus stays fresh, and failed videos are never billed.
videoId
Optional
YouTube video ID, e.g. jNQXAC9IVRw. Null on error items where the input URL could not be resolved to a video.
url
Optional
Canonical video URL.
title
Optional
Video title.
channel
Optional
Channel display name.
channelId
Optional
YouTube channel ID.
uploadDate
Optional
Video upload date as YouTube's own absolute date string (hl=en, e.g. "Jul 11, 2018"), taken from the same next call as chapters. Null when includeChapters is disabled or the date is unavailable.
durationSeconds
Optional
Video duration in seconds.
language
Optional
Language tag of the delivered caption track.
matchedLanguageTag
Optional
The track's BCP-47 language tag (e.g. 'en', 'pt-BR'). Under allLanguages, if a video has two tracks of the same language and kind, the second carries a '#2'/'#3' suffix so each item stays distinct — use the separate 'language' field for the clean code.
isAutoGenerated
Optional
Whether the caption track is auto-generated (ASR).
availableLanguages
Optional
Every caption track actually available for the video (manual + auto-generated; never translation targets).
sourceUrls
Optional
The input URL(s) that led to this video (video, playlist, or channel URLs).
sourcePosition
Optional
Zero-based position of the video within its expanded source (playlist/channel).
segments
Optional
Raw caption cues in caption order, each with millisecond timing. The RAG-ready paragraphs live inside chapters[].paragraphs; these are the verbatim cues they were built from.
chapters
Optional
Chapter-grouped, chunk-ready paragraphs (the RAG wedge). Always present on a delivered transcript: when a video has no chapters (or includeChapters is off) the whole transcript is one group with a null title. Caption text before the first chapter is a leading null-title group.
plainText
Optional
Full transcript as plain text. Present when formats includes "text".
srt
Optional
Transcript rendered as SRT. Present when formats includes "srt".
vtt
Optional
Transcript rendered as WebVTT. Present when formats includes "vtt".
errorReason
Optional
Machine-readable error category on error items (e.g. noCaptions, videoUnavailable). Null on successful items.
rawReason
Optional
Raw upstream error detail, for debugging. Null on successful items.
input
Optional
Echo of the offending input context on error items. Null on successful items.
