YouTube Search Scraper: Videos & Channel Discovery

YouTube Search Scraper

YouTube search scraper and YouTube search API: run a YouTube video search by keyword and get one row per result, in the order YouTube ranks them: videos, Shorts, channels, playlists or movies, with views, age, length, the channel (name, handle, avatar, verified badge), badges, chapters and the description snippet. Search from any country and in any language, with YouTube's own filters. Exact likes, comments, publish time, tags and channel stats are an optional add-on. No login, no YouTube API key, no browser.

#	🎬 Video	📺 Channel	👀 Views	📅 Age	⏱️ Length	🏅 Badges	🔎 Search term
1	Python Full Course for Beginners	Programming with Mosh ✅	49,465,354	7 years ago	06:14:07	CC	python tutorial
2	Python Full Course for free 🐍	Bro Code ✅	13,170,298	2 years ago	12:00:00	Fundraiser	python tutorial
3	Learn Python for Beginners - Visually Explained	Visually Explained ✅	183,005	1 month ago	03:54:33	–	python tutorial
1	I Tested NYC Street Food in Every Borough	Jon Barr ✅	404,137	5 months ago	00:23:25	4K	street food
2	$2 Burgers in Harlem | Street Food Icons	Munchies ✅	24,011,121	7 years ago	00:09:12	CC	street food
3	New York Street Food Marathon!! From $1 to $1000!!	More Best Ever Food Review Show ✅	3,022,812	5 months ago	01:24:34	4K	street food

$0.20 per 1,000 results on Starter ($0.24 Free, $0.17 Scale, $0.15 Business). The prefilled run (100 results) costs about $0.02. Nothing per run. Full details add $0.10 per 1,000 results, only when you turn them on.

Plan	Per 1,000 results	With full details
Free	$0.24	$0.34
Starter	$0.20	$0.30
Scale	$0.17	$0.27
Business	$0.15	$0.25

What you type

Input	Example	What it does
🔎 Search terms	python tutorial, street food, or a link like https://www.youtube.com/results?search_query=lofi+hip+hop&sp=CAMSAhAB	One YouTube search per term (a link keeps its own filters)
🏷️ Result type	Videos, Shorts, Channels, Playlists, Movies	YouTube's Type filter: one type per run
🔢 Results per search term	50 (100 when empty, up to 5,000)	How deep each search goes
➕ Full details	on / off (off)	Adds the watch page's or the channel's About panel's fields (below)
🎛️ Sort by, upload date, duration, features	Popularity, This week, Under 3 minutes, HD + Subtitles	YouTube's own search filters, with the same values as its filter dialog
🌍 Country, 🗣️ Language	MX, es	Search as from that country in that language: YouTube returns different results (es-MX shared 5 to 7 of the first 20 with en-US in our tests)

A result that several of your terms find is saved and charged once. Ages and counts are read in YouTube's languages (Spanish hace 21 horas, German vor 5 Tagen, Japanese 14 時間前, Korean 구독자 177만명, Hindi 17.7 लाख...), so every row has numbers as numbers and the text as YouTube wrote it.

What you get

Result type	Fields
🎬 Videos, Shorts, movies	position (rank in its search), title, type (video, shorts, stream, movie), id, url, thumbnailUrl, thumbnails, previewUrl (the moving preview YouTube plays on hover, a 6-second animated WebP), richThumbnail (its sizes), viewCount (exact), viewCountText, date (the day the age gives), dateText (the age as shown), dateEstimated, duration, durationSeconds, text (the description snippet), badges (New, 4K, CC, Free with ads...), isLiveNow, streamType (live, scheduled, ended), scheduledStartAt, isMembersOnly, chapters (title, start, startSeconds, thumbnailUrl), channelName, channelId, channelUrl, channelUsername, channelAvatarUrl, isChannelVerified, isChannelArtist; movies also movieMetadata (genre and year, actors, directors)
📺 Channels	position, title (the channel's name), id, url, thumbnailUrl (avatar), channelUsername (handle), numberOfSubscribers, subscribersText, text (description), isChannelVerified, isChannelArtist
📚 Playlists	position, title, id, url, thumbnailUrl, playlistType (playlist, course, podcast...), videoCount, firstVideoId, firstVideos (the videos the search lists under it), text, channelName, channelUrl, channelUsername, isChannelVerified
🧾 Every row	searchQuery, input, fromYTUrl (the search with its filters), country, language, scrapedAt

➕ With full details, video, Short and movie rows also get likes (exact), commentsCount (exact), commentsTurnedOff, the exact publish time in date, the full description in text, descriptionLinks, hashtags, keywords (tags), category, originalLanguage and originalLanguageCode (the language spoken in the video), audioTracks (dubbed videos: each audio track's language, and whether it's the original or YouTube's automatic dub), location, isAgeRestricted, isUnlisted, isPaidContent ("Includes paid promotion"), isMonetized (YouTube serves ads on it to a logged-out viewer; null on Shorts), liveStartedAt, liveEndedAt, Shorts' soundTitle, musicTitle, musicArtist, collaborators, and the channel's numberOfSubscribers, channelDescription, channelDescriptionLinks, channelJoinedDate, channelCreatedAt (when the run read the channel's feed for a publish time), channelLocation, channelTotalVideos, channelTotalViews. Channel rows get the About panel: channelTotalViews, channelTotalVideos, channelJoinedDate, channelLocation, channelDescription, channelDescriptionLinks, and channelCreatedAt (the channel's creation time to the second, from its RSS feed; null when YouTube's feed doesn't answer). Playlists are always saved as the search shows them.

The dataset has five views: Search results, Video details, Full details (videos), Channels and Playlists.

Measured on 16 runs on Apify on 2026-10-17 (4,135 video rows, 200 channel rows in es-MX, 60 playlists, 150 Shorts, 40 movies, 140 rows with full details):

Field	Filled
title, id, url, thumbnailUrl, viewCount, channelName, channelId (videos)	100%
date, dateText, duration (videos)	100%
text (videos' snippet)	100%
badges (videos)	59%
chapters (videos)	38%
channelUsername (videos)	93%
numberOfSubscribers (channels)	96%
videoCount, firstVideos (playlists)	100%
channelId, viewCount (Shorts)	100%
likes, commentsCount, category (full details, videos)	100%
keywords (full details, videos)	88%
channelJoinedDate, channelTotalViews (full details)	97-100%

badges and chapters are filled where YouTube shows them; channelUsername is empty for channels without a handle; a channel can hide its subscriber count. Most Shorts searches answer as YouTube's Shorts grid, which shows only the title, views (rounded, 1.2M) and the channel's id: turn on full details for a Short's channel name, exact views, likes, comments, date, length and sound. Movies show no views or age in search.

Moving from another YouTube search Actor

Paste your saved input as it is: the fields below are read the way those Actors read them.

Their input	Read here as
api-ninja: query, maxResults, type, sortBy, duration, uploadDate, features, geo, lang, scrapeAllResults, videoDepthDetails	the same search; standard / detailed / complete depth = full details
ninhothedev: queries, query, maxItems, maxResults	maxItems per query, maxResults as the total across queries, in input order ({"queries": ["python tutorial", "javascript tutorial"], "maxItems": 10, "maxResults": 15} saves 10 + 5); a non-empty queries wins over query
grow_media: q, maxResults, order, videoDuration, regionCode, relevanceLanguage, useFilters	the same search; with useFilters: false its filters are left out, as it does
grow_media: publishedAfter, publishedBefore	YouTube's after: / before: search operators, to the day (fromYTUrl shows them)
grow_media: eventType	live: YouTube's Live filter; completed and upcoming: YouTube's search has no filter for them, so streams with that label are kept from its results (upcoming streams are rarely in a search)
grow_media: videoDefinition: high, videoLicense: creativeCommon, safeSearch: strict	YouTube's HD and Creative Commons filters, Restricted Mode

Not available on YouTube's search, so searched without them (the run's notices say so): videoDefinition: standard, videoLicense: youtube, order: title / videoCount, channelType: show; safeSearch: moderate is YouTube's normal search. Search links in startUrls keep their own filters. One difference: a video that several of your search terms find is saved and charged once (under the first term), where ninhothedev saves it once per query.

Their field	Here
videoId, video_id	id
channelTitle, channel	channelName
description	text
publishedTimeText, published_time	dateText
publishDate, publishedAt	date
lengthText, duration_seconds	duration, durationSeconds
thumbnail	thumbnails (thumbnailUrl = the largest)
richThumbnail	richThumbnail (previewUrl = the largest)
views, views_text	viewCount, viewCountText
rank, order	position
query	searchQuery
channelCreatedAt, channelVideoCount, channelViewCount	channelCreatedAt (or the day in channelJoinedDate), channelTotalVideos, channelTotalViews (full details)

	api-ninja/youtube-search-scraper	grow_media/youtube-search-api	ninhothedev/youtube-search-scraper	This Actor
Price per 1,000 results, paid plans	$0.25 until 2026-10-26, then $0.50, plus $0.01 per run	$0.60 / $0.50 / $0.40	$0.50	$0.20 / $0.17 / $0.15, no start fee
Result types	videos, Shorts, channels, playlists, movies	videos, channels, playlists	videos	videos, Shorts, channels, playlists, movies
Country and language	yes	yes	no	yes
Likes, comments, channel stats	deeper modes	yes	no	full details (+$0.10 per 1,000)
Chapters	key moments in its deeper modes	no	no	yes, in search rows
Badges, live and upcoming streams	not declared	live/upcoming filter	no	yes

Prices are from the Apify Store on 2026-10-16.

Output

One row per result. A real video row, shortened:

{
  "position": 5,
  "title": "iPhone 17 After 1 Year: Why I Still Love It",
  "type": "video",
  "id": "S71ljmbtERg",
  "url": "https://www.youtube.com/watch?v=S71ljmbtERg",
  "thumbnailUrl": "https://i.ytimg.com/vi/S71ljmbtERg/hq720.jpg",
  "viewCount": 71252,
  "viewCountText": "71,252 views",
  "date": "2026-10-03",
  "dateText": "2 weeks ago",
  "dateEstimated": true,
  "duration": "00:11:36",
  "durationSeconds": 696,
  "text": "iPhone 17 1 year review, iPhone 17 review, Apple iPhone 17 camera test, iPhone 17 battery life, iPhone 17 vs iPhone 17 Pro, ...",
  "badges": ["1 product", "Earns commission"],
  "isLiveNow": false,
  "isMembersOnly": false,
  "chapters": [
    {"title": "Design", "start": "00:00:00", "startSeconds": 0, "thumbnailUrl": "https://i.ytimg.com/vi/S71ljmbtERg/hqdefault_25233.jpg"},
    {"title": "Display", "start": "00:01:41", "startSeconds": 101, "thumbnailUrl": "https://i.ytimg.com/vi/S71ljmbtERg/hqdefault_102866.jpg"}
  ],
  "channelName": "Everyday Legends",
  "channelId": "UCbM-fdGZyGd-NNRpYscDeHA",
  "channelUrl": "https://www.youtube.com/@Everyday_Legends",
  "channelUsername": "Everyday_Legends",
  "isChannelVerified": false,
  "searchQuery": "iphone 17 review",
  "fromYTUrl": "https://www.youtube.com/results?search_query=iphone+17+review&sp=EgIQAQ%3D%3D",
  "country": "US",
  "language": "en",
  "scrapedAt": "2026-10-17T02:18:37Z"
}

A channel search in Mexico, in Spanish:

#	📺 Channel	👤 Handle	👥 Subscribers	✅ Verified
1	El Pan De Valdo	@ElPanDeValdo	234,000 (234 K suscriptores)	yes
2	COMO HACER PAN SV	@comohacerpanSV	59,000 (59 K suscriptores)	no
3	Mirecetadepan.com	@Mirecetadepancom	309,000 (309 K suscriptores)	no

With full details:

🎬 Video	👀 Views	❤️ Likes	💬 Comments	📅 Published	🗂️ Category	👥 Subscribers	📍 Country
Learn Python - Full Course for Beginners [Tutorial]	49,468,302	1,123,445	46,093	2018-07-11T18:00:42Z	Education	11,900,000	United States
Python Full Course for Beginners	49,373,083	1,273,475	62,928	2019-02-18T15:00:08Z	Education	5,170,000	United States

The run's summary (key-value store record OUTPUT) lists every search term with how many results it gave and its status, the result type, country, language and filters, and repeats skipped.

From a keyword to videos, comments and transcripts

A YouTube keyword scraper is the first step of most research. This is how the Actors chain:

1. Search a keyword here (videos, Shorts, channels or playlists, one type per run). Basic rows are the cheapest way to list what YouTube shows for a term; turn on Full details when you need exact likes, comments, tags and channel stats in the same rows.
2. Videos: the url of each row goes into YouTube Video Details Scraper (full metadata), YouTube Comments Scraper (comments and replies) or YouTube Transcript API (what is said, with timestamps).
3. Channels: run the search with the type Channels, then paste the url of the channels you like into YouTube Channel Scraper for all their videos. To filter creators by subscribers, country or verification, use YouTube Channel Search Scraper.
4. Playlists: paste a playlist row's url into YouTube Playlist Scraper.

Searching and reading videos in one run? YouTube Scraper takes search terms and returns full video rows.

Did it help?

A short review helps others find it: https://apify.com/deepmine/youtube-search-scraper/reviews