YouTube Video Scraper — video, channel, comments, transcript and search in one row

Paste one YouTube video URL and get five joined datasets in a single row: video metadata, the channel profile behind it, the comments, the full transcript, and the YouTube search results for the video's own title — with the rank the video holds on that query.

Built for agencies and competitive-intelligence teams who today run three or four separate actors and glue the exports together by hand. Here every section arrives in one flat row, joined on video_id and channel_id, at $0.60 per delivered video on the Free plan — from the Console, the API, the CLI or the Apify MCP server (mcp.apify.com).

What does the YouTube Video Scraper do?

For each unique video it fetches up to five sections in one pass: video (title, counters, duration, keywords, key moments), channel (handle, subscribers, lifetime views, video count, join date), comments (top-level comments with likes and reply counts), transcript (timestamped segments plus one plain-text string) and search (the results page for the video's own title, and where the video ranks on it).

Every section lands in the same row, always with the same 36 keys, so the warehouse join is one line instead of four exports with four field naming schemes. Switch a section off and it is not fetched at all: its fields come back null and the section is listed in sections_skipped. Nothing is summarised, translated or scored — the row is what YouTube published, measured at one moment.

Why use the YouTube Video Scraper

• Five datasets, one run, one join key — one row joined on video_id and channel_id, so no cross-actor field mapping.
• Competitive rank included — search_self_rank shows where the video sits on the results page for its own title, next to the competing videos in search_results.
• Duplicates merged before anything is fetched — the same video pasted three ways (watch?v=, youtu.be/, bare ID) is one row and one charge.
• Missing data is reported, not faked — a video without subtitles is a delivered row with status: "partial" and transcript_unavailable in partial_reasons, not a failed run.
• Free repeats for 24 hours — a video processed in the last 24 hours is re-delivered from cache at no charge, keeping its original fetched_at.
• A hard spending limit you set — maxTotalChargeUsd stops the run before its first request.

How to scrape a YouTube video, its channel, comments and transcript

1. Paste links into Video URLs or IDs — watch URLs, youtu.be links, Shorts links, embed links and bare 11-character IDs all work.
2. Leave the five Sections switches on for the full pack, or turn off what you do not need.
3. Pick the Language for the transcript (English by default, 28 languages).
4. Optionally pick a Country to request the data from one market; "Auto" lets YouTube decide.
5. Raise Comment pages per video (about 20 top-level comments per page, up to 10 pages).
6. Optionally set Maximum charge per run (USD) as a hard ceiling.
7. Run the Actor, then open the Overview dataset view or export to JSON, CSV or Excel. The Issues view lists only rows that came back partial or not_found.

Minimal input:

{ "videoUrls": ["https://www.youtube.com/watch?v=rfscVS0vtbw"] }

Transcript and comments only, Spanish, two comment pages, capped budget:

{ "videoUrls": ["rfscVS0vtbw", "https://youtu.be/UF8uR6Z6KLc"], "includeChannel": false,
  "includeSearch": false, "language": "es", "commentsPages": 2, "maxTotalChargeUsd": 0.05 }

What data you get per video

One row = one video, always the same 36 keys. Example row, long lists truncated:

{
  "video_id": "rfscVS0vtbw",
  "video_url": "https://www.youtube.com/watch?v=rfscVS0vtbw",
  "status": "ok",
  "fetched_at": "2026-07-30T09:14:02.371Z",
  "title": "Learn Python - Full Course for Beginners [Tutorial]",
  "description": "This course will give you a full introduction into the core concepts ...",
  "views": 49046089, "likes": 1100000, "published_date": "Jul 11, 2018", "duration_seconds": 16012,
  "keywords": ["python", "python tutorial", "python full course"],
  "thumbnail": "https://i.ytimg.com/vi/rfscVS0vtbw/maxresdefault.jpg",
  "comments_total": 46111, "chapters": [],
  "key_moments": [{ "start": "1:45", "title": "Installing Python & PyCharm" }],
  "channel_id": "UCWv7vMbMWH4-V0ZXdmDpPBA", "channel_handle": "@freecodecamp",
  "channel_title": "freeCodeCamp.org", "channel_subscribers": 11800000,
  "channel_views": 1252445166, "channel_videos_count": 1900, "channel_joined_date": "2014-12-16",
  "channel_description": "We're a community of millions of people learning to code ...",
  "comments": [
    { "comment_id": "Ugxqcek3RUzxWv-4VqB4AaABAg", "author": "@freecodecamp", "text": "Want more from Mike?",
      "likes": 4700, "replies": 147, "published_date": "5 years ago" }
  ],
  "transcript": [{ "text": "Welcome to this Python course.", "start": 1.36, "duration": 3.2, "end": 4.56, "lang": "en" }],
  "transcript_text": "Welcome to this Python course. In this video we will cover ...",
  "transcript_language": "en", "transcript_segment_count": 4212, "transcript_word_count": 61840,
  "search_query": "Learn Python - Full Course for Beginners [Tutorial]",
  "search_results": [
    { "position": 1, "video_id": "UF8uR6Z6KLc", "title": "Python for Everybody",
      "url": "https://www.youtube.com/watch?v=UF8uR6Z6KLc", "channel_name": "freeCodeCamp.org",
      "channel_handle": "@freecodecamp", "views": 1200000, "published_date": "5 years ago", "duration": "13:39:11" }
  ],
  "search_self_rank": 2,
  "sections_included": ["video", "channel", "comments", "transcript", "search"],
  "sections_skipped": [], "partial_reasons": [],
  "source": { "video": "primary", "channel": "primary", "comments": "primary", "transcript": "primary", "search": "primary" }
}

Field	Type	What it tells you
video_id / video_url	string	Canonical 11-character ID and watch URL — the join key
status / fetched_at	string	ok, partial or not_found, and when the row was produced
title / description	string | null	Title and description, decoded to plain text
views / likes	integer | null	View and like counts as integers
published_date / duration_seconds	string | integer	Publish label as listed, length in seconds
keywords / thumbnail	array | string	Keyword tags, highest-resolution thumbnail URL
comments_total	integer | null	Exact comment count on the video
key_moments / chapters	array	Timeline markers {start, title}; chapters is the structured field, often empty
channel_id / channel_handle / channel_title	string | null	UC… ID, @handle, name — second join key
channel_subscribers / channel_views / channel_videos_count	integer | null	Channel counters
channel_joined_date / channel_description	string | null	Creation date and About text
comments	array | null	{comment_id, author, text, likes, replies, published_date}
transcript	array | null	{text, start, end, duration, lang} in seconds, end = start + duration
transcript_text / transcript_segment_count / transcript_word_count	string | integer	Whole transcript as plain text (entities and markup removed), plus its counters
transcript_language	string | null	Language actually returned, never an echo of your request
search_query / search_results	string | array	The video's own title, and the results page for it
search_self_rank	integer | null	Rank of this video on that page; null when it is not on it
sections_included / sections_skipped	array	Which of the five sections delivered data, and which did not
partial_reasons	array	Fixed codes: transcript_unavailable, channel_unavailable, video_not_found, …
source	object	Per section: primary, fallback, or null when there was no response

Input parameters

Field	Required	Notes
videoUrls	yes	Up to 500 video URLs or IDs; duplicates merge into one row
includeVideo	no	Video metadata section, on by default
includeChannel	no	Channel profile section, on by default
includeComments	no	Comments section, on by default
includeTranscript	no	Transcript section, on by default
includeSearch	no	Search + search_self_rank section, on by default
language	no	Transcript language, 28 options, en by default
country	no	Market the data is requested from; "Auto" lets YouTube choose
commentsPages	no	Comment pages per video, 1-10 (default 1, ~20 comments per page)
maxTotalChargeUsd	no	Hard ceiling; the run stops before its first request

How much does the YouTube Video Scraper cost?

Pay-per-event billing: you pay per delivered row, never per request or per section.

Event	When	Price
Run start	Once per run	$0.002
YouTube video report	Per delivered video	$0.60 Free → $0.10 Bronze → $0.08 Silver → $0.06 Gold+

The whole pack is one price: five sections, up to 10 comment pages and retries sit inside the per-video price, and turning sections off does not change it. 100 videos cost $60.20 on the Free plan, $10.20 or less on any paid plan.

You are not charged for cache hits (a video processed within 24 hours is re-delivered free), for failures (no row, no charge), or for duplicate URLs of the same video. A not_found video is charged as one row — "this video does not exist" is a delivered answer. A partial row is one row too: measured data plus an explicit list of what was missing.

YouTube Video Scraper vs single-purpose YouTube actors

	This actor	Separate metadata / comments / transcript actors
Input	One video URL	The same URL pasted into 3-4 actors
Datasets per run	5, in one row	3-4 exports, 3-4 schemas
Join key	video_id + channel_id	Whatever each actor calls its ID
Rank for the video's title	Included	Usually another actor
Billing unit	One delivered video	One per actor, per model

What this actor does not do

• It does not summarise, translate, score or sentiment-tag anything.
• It does not promise a filled chapters array: YouTube often exposes the chapter list only as key_moments, and nothing is copied between the two fields.
• It does not return a commenter's display name (author is the handle the comment carries), and it returns top-level comments only, with a replies count.
• It does not list a channel's back catalogue, its Shorts or its upload cadence.
• It does not touch private, members-only or age-gated content, downloads no media, and does not go past 500 videos per run.

FAQ

What exactly is one billed row?

One unique video that came back with data — five sections or one, it is one charge. Duplicates, cache hits and failures are not charged.

What happens to a video without a transcript?

You get a normal row with status: "partial", null transcript fields, transcript in sections_skipped and transcript_unavailable in partial_reasons. The other four sections are still delivered, so a captionless video never breaks a batch.

Why is search_self_rank sometimes null?

Because the video is not on the returned results page for its own title. null means "not on this page" — the field never returns 0 or a negative number to mean "missing".

Can I get more than 20 comments per video?

Yes. Set Comment pages per video up to 10; every page lands in the same comments array of the same row, at no extra charge.

Does the requested language guarantee that transcript language?

No, and the row says so. transcript_language reports the language actually returned; a video with no subtitles in your language yields transcript_unavailable, not a silent substitution.

Can I run this from an AI agent or on a schedule?

Yes — Apify API, CLI, or the Apify MCP server (mcp.apify.com). Schedule runs to track how counters, comments and search rank move over time.

Related actors

• YouTube Transcript Scraper — transcripts for a whole channel or playlist.
• YouTube Influencer Finder — find channels worth analysing, with subscriber stats and public contact profiles.
• YouTube Research Digest — an AI digest of what the top videos for a keyword say.

Legal & support

• This actor collects publicly available video information: what a video page, a channel page, a public comment thread, a published subtitle track and a results page show to any visitor. It does not access private, members-only or log-in-gated content and downloads no media files.
• Comments and transcripts are returned verbatim, with timestamps as published, so every line can be verified against the source video.
• Video metadata, comments and transcripts are third-party content. You remain responsible for how you use them, including copyright, YouTube's terms and any attribution or storage rules that apply to you — especially when this data feeds a public product or an AI index.
• Comment rows can contain personal data (handles, free text). Handle them under your own GDPR or equivalent obligations and keep only what your purpose requires.
• Output is as-is: every counter and ranking is a point-in-time value.
• Questions or a bug? Open an issue on the Actor's Issues tab, or see the API tab.

SEO keywords

youtube scraper, youtube video scraper, youtube data api alternative, scrape youtube video data, youtube comments scraper, youtube transcript scraper, youtube channel scraper, youtube search scraper, youtube metadata extraction, youtube video statistics, youtube subscriber count api, youtube competitor analysis, youtube seo tool, youtube ranking checker, video engagement data, youtube comment export, youtube captions api, bulk youtube data extraction, youtube dataset json csv, youtube influencer research.

Built by doesaiknow.