YouTube Intelligence - Metadata, Transcript, Chapters, Entities
Pricing
Pay per usage
Bulk YouTube video data: metadata (title, channel, views, likes, comments, thumbnails) + timestamped transcript + auto-detected chapters + extracted entities/keywords/hashtags + optional auto-translation. Pay only for what you use - 3 tiers.
input
Required
Original URL or ID supplied as input.
video_id
Required
11-character YouTube video ID.
scraped_at
Required
tier
Required
available
Required
True if metadata could be fetched.
reason
Optional
Failure reason when available=false.
title
Optional
description
Optional
channel_id
Optional
channel_name
Optional
channel_handle
Optional
channel_url
Optional
subscriber_count
Optional
duration_seconds
Optional
view_count
Optional
like_count
Optional
comment_count
Optional
upload_date
Optional
YYYY-MM-DD.
publish_date
Optional
YYYYMMDD raw form.
live_status
Optional
e.g. 'not_live', 'is_live', 'was_live'.
is_short
Optional
categories
Optional
language
Optional
age_limit
Optional
thumbnail_url
Optional
thumbnails
Optional
chapters
Optional
Auto-detected chapters from YouTube or parsed from description timestamps.
chapters_source
Optional
mentions
Optional
description_links
Optional
availability
Optional
webpage_url
Optional
transcript
Optional
Transcript data (only on enriched/premium tier).
keywords
Optional
Top transcript keywords by frequency (enriched/premium tier).
named_entities
Optional
DEPRECATED v0.2 — kept for back-compat. Use named_entities_categorized instead.
named_entities_categorized
Optional
Categorized heuristic named-entity extraction over the transcript.
transcript_translated
Optional
Translated transcript text (premium tier with translate_to set).
engagement_metrics
Optional
Engagement ratios and view velocity (per-day, per-hour, trending flag).
seo_metrics
Optional
Title/description/tag analytics: length, hashtag counts, keyword density, CTA detection, timestamp count.
content_type_classification
Optional
Heuristic classification of the video into one of 20 content types with top-3 confidence scores.
sponsorship_detected
Optional
Sponsor-read phrase detection in the transcript with approximate positions.
speaking_pace_wpm
Optional
Words per minute computed from transcript segments (enriched/premium tier).
filler_word_rate
Optional
Percentage of transcript tokens that are filler words (um, uh, like, you know, ...).
transcript_urls
Optional
URLs spoken/mentioned in the transcript text (separate from description_links).
comments_top_5
Optional
Convenience field: the top 5 fetched comments ranked by like_count.
comment_sentiment
Optional
Lexicon-based sentiment over fetched comments.
channel
Optional
Channel-level enrichment. Cached per channel across the run. Only present when fetch_channel_data=true.
comments
Optional
Top N comments with author, likes, replies, pinned/hearted flags. Only present when fetch_comments=true.