YouTube Scraper – Videos, Shorts, Channels & Search

YouTube Scraper – Videos, Shorts, Channels & Search — complete video data from searches, channels and playlists (2026)

Get YouTube videos, Shorts and live streams as data from search terms or YouTube links: title, link, exact publish date and time, duration, views, likes, comment count, tags, category and channel data, and optionally the comments. No YouTube account, cookies or API key are needed. You pay only for the videos you get, $2.50 per 1,000 videos (less on higher Apify plans).

YouTube is the largest video site, so its titles, views and comments are the raw material for creator research, brand monitoring, content planning and AI pipelines. This Actor is an alternative to the YouTube Data API for public data.

Why use this YouTube scraper

• Reliable, measured on the cloud: 15 runs out of 15 succeeded in our tests of October 20, 2026 (5 inputs, each run 3 times, 453 videos, no duplicates).
• Complete rows: title, link, views, likes, channel name and subscribers were filled on 94-100% of the videos in those runs. Duration and comment count are empty only on live streams that are still running; tags, category and exact date need the default "Full details" option.
• Fast: the example input gives 20 videos in 4.6-12.3 seconds.
• Newest first and Most viewed. YouTube no longer sorts search results by date and sorts by views only roughly, so the Actor reads the result list of the search and sorts it itself. For Newest first it also merges YouTube's own "last hour", "today" and "this week" lists, which gave 154-189 videos of the last 3 days in our tests, against 4-21 from the plain list. YouTube still changes what it lists from one request to the next: two runs a few minutes apart shared 26-33 of the 40 newest videos.
• Links from the description. descriptionLinks lists the links written in each video description with their real address (sponsors, shops, social pages), for creator and sponsor research.
• Only new videos. publishedAfter (2026-09-01 or 7 days) keeps only recent videos; on channels sorted newest first the Actor stops at the first older video. Older videos are not charged.
• Exact comment count and big channels. Where YouTube shows a rounded "2.4K comments", you get the exact number. Playlists go past the first 100 videos.
• Cheaper than the market leader. $2.50 per 1,000 videos on the Free plan, against $4.00 for the most used YouTube scraper on the Apify Store (checked on October 20, 2026).
• You pay only for videos. Searches with no results, wrong links and refused pages are free status rows. Duplicates are removed and not charged.

How to use YouTube Scraper – Videos, Shorts, Channels & Search

1. Fill in Search terms (searchQueries) or YouTube URLs (startUrls).
2. Set Max videos per search or URL (maxResults, default 20) and Comments per video (maxComments, default 0) and Videos in parallel (maxConcurrency, default 20).
3. Click Start. The example input takes 4.6-12.3 seconds.
4. Download the results as JSON, CSV, Excel or HTML, or read them through the Apify API.
5. To follow a channel or a keyword over time, save the input as a task, add publishedAfter (for example 1 day) and schedule it: each run returns only the videos published since.

Input

Field	Description
Search terms (searchQueries)	One search per line, typed as you would in the YouTube search bar. Use the filters below to narrow the results.
YouTube URLs (startUrls)	Links to videos, Shorts, channels (/@handle, /channel/UC…, /c/…, /user/…, optionally ending in /videos, /shorts or /streams), playlists, search result pages (their filters are kept) and hashtag pages (/hashtag/cooking).
Max videos per search or URL (maxResults)	For channels, the limit applies to each tab you choose (Videos, Shorts, Live). Default 20.
Full details for every video (fullDetails)	Exact publish date and time, duration, tags and category for every row (also for search, channel and playlist results). Turn off only if you need speed over completeness: views, likes, comment count, date, description and channel data are included either way. Default true.
Search: sort by (sortBy)	Relevance keeps the YouTube order. YouTube no longer sorts search results by date and sorts by views only roughly, so Newest first and Most viewed read the result list (up to 1,000 videos) and sort it here; Newest first also merges YouTube's last hour, today and this week lists. YouTube changes its lists between requests, so two runs minutes apart share most, not all, of the newest videos. Default "relevance".
Search: upload date (uploadDate)	Only videos uploaded in this period (YouTube filter). Default "any".
Search: type (videoType)	Videos, Shorts or live streams. "Everything" returns what the YouTube page shows, Shorts shelves included. Default "all".
Search: duration (duration)	Video length (YouTube filter). Default "any".
Search: features (features)	YouTube search features: all selected must match. Default [].
Channel tabs (channelTabs)	Which tabs to read from channel URLs (a URL ending in /videos, /shorts or /streams reads only that tab). Default ["videos"].
Channel order (channelSort)	Same orders as the channel page. With newest first, publishedAfter stops the scraper at the first older video. Default "newest".
Only videos published after (publishedAfter)	A date (2026-09-01) or an age (7 days, 2 weeks, 3 months). Works for searches, channels, playlists and videos; on channels sorted newest first the scraper stops as soon as it reaches older videos. Older videos are not charged.
Comments per video (maxComments)	0 = no comments. Comments are added to each video row (text, author, likes, replies, pinned, hearted by the creator) and charged per comment. Default 0.
Comments order (commentsSort)	Top comments or newest first, as on YouTube. Default "top".
Country (country)	Two-letter country code (US, GB, DE, IN, BR…): YouTube ranks search results and checks video availability for this country. Default "US".
Search language (language)	Language code for search results (en, es, de, pt, fr, hi…). Used with Relevance order; Newest first and Most viewed read the result list in English to sort it. Default "en".
Videos in parallel (maxConcurrency)	How many videos are read at the same time. Higher is faster; lower is gentler on YouTube. The default of 20 suits most runs. Default 20.

Example input (20 videos in 4.6-12.3 seconds in our runs of October 20, 2026):

{
  "searchQueries": [
    "lofi hip hop"
  ]
}

Output example

One row per video. This is a real row from a cloud run of October 20, 2026:

{
  "type": "short",
  "id": "IHkp5Kq72ro",
  "url": "https://www.youtube.com/shorts/IHkp5Kq72ro",
  "title": "How Dangerous Are Non-Stick Pans?",
  "description": "What happens to your body if you eat Teflon from a non-stick pan?\n\n#pfas \n#foreverchemicals",
  "publishedAt": "2026-10-01T18:26:42.000Z",
  "publishedDate": "2026-10-01",
  "publishedTimeText": "2 weeks ago",
  "durationSeconds": 101,
  "duration": "1:41",
  "viewCount": 1626785,
  "likeCount": 47243,
  "commentCount": 1037,
  "commentsTurnedOff": false,
  "hashtags": [
    "#pfas",
    "#foreverchemicals"
  ],
  "tags": [
    "veritasium",
    "science",
    "physics",
    "pfas"
  ],
  "category": "Education",
  "thumbnailUrl": "https://i.ytimg.com/vi/IHkp5Kq72ro/hqdefault.jpg",
  "isLive": false,
  "isUpcoming": false,
  "liveViewers": null,
  "liveStartedAt": null,
  "liveEndedAt": null,
  "isFamilySafe": true,
  "isUnlisted": false,
  "isMembersOnly": false,
  "channelId": "UCHnyfMqiRRG1u-2MsSQLbXA",
  "channelName": "Veritasium",
  "channelHandle": "@veritasium",
  "channelUrl": "https://www.youtube.com/@veritasium",
  "channelSubscribers": 21400000,
  "channelVerified": true,
  "collaborators": null,
  "channelDescription": "An element of truth - videos about science, education, and anything else we find interesting.",
  "channelCountry": "United States",
  "channelJoinedDate": "2010-07-21",
  "channelTotalViews": 4680895501,
  "channelTotalVideos": 541,
  "channelLinks": [
    {
      "title": "Elements of Truth - The Game",
      "url": "https://ve42.co/YTBio"
    },
    {
      "title": "Buy SNATOMS",
      "url": "https://snatoms.com/"
    },
    {
      "title": "Facebook",
      "url": "http://facebook.com/veritasium"
    },
    {
      "title": "TikTok",
      "url": "https://www.tiktok.com/@veritasium?lang=en"
    },
    {
      "title": "Twitter",
      "url": "http://twitter.com/veritasium"
    },
    {
      "title": "Instagram",
      "url": "https://www.instagram.com/veritasium_daily/"
    }
  ],
  "source": "channel",
  "input": "https://www.youtube.com/@veritasium",
  "position": 3,
  "detailsComplete": true,
  "scrapedAt": "2026-10-20T10:15:16.975Z"
}

Field	What it is
thumbnailUrl	Link to the video thumbnail image
title	Video title as published
type	video, short or stream (live now, upcoming or past live stream)
url	Link to the video or Short on YouTube
publishedAt	Exact publish date and time (ISO 8601, UTC); empty when full details are off and the list does not show it (filled on about 84% of the videos in our tests)
publishedDate	Publish date as YYYY-MM-DD
duration	Video length as hours:minutes:seconds (filled on about 80% of the videos in our tests)
viewCount	Exact number of views; empty for members-only videos
likeCount	Number of likes; empty when the creator hides likes
commentCount	Exact number of comments; empty when comments are turned off (filled on about 80% of the videos in our tests)
channelName	Channel name
channelSubscribers	Subscribers, rounded as YouTube shows them (21.3M becomes 21300000); empty when hidden
category	YouTube category chosen by the creator (full details only) (filled on about 84% of the videos in our tests)
input	The search term or link of your input that produced this row
position	Position of the video in its search, channel or playlist, starting at 1
id	YouTube video id (the v= part of the link)
description	Full video description as published
descriptionLinks	Links written in the video description, in order and without repeats: the text as shown and the real address (YouTube redirects are resolved); empty list when there are none (filled on about 55% of the videos in our tests)
publishedTimeText	Publish time as YouTube shows it on the page, for example 7 days ago
durationSeconds	Video length in seconds; empty while a live stream is running (filled on about 80% of the videos in our tests)
commentsTurnedOff	true when the creator turned comments off for this video
hashtags	Hashtags written in the title or description (filled on about 47% of the videos in our tests)
tags	Tags set by the creator (full details only) (filled on about 64% of the videos in our tests)
isLive	true while the video is a live stream that is running
isUpcoming	true for a scheduled live stream or premiere that has not started
liveViewers	People watching right now, only while a live stream is running (filled on about 20% of the videos in our tests)
liveStartedAt	When the live stream started (ISO 8601, UTC); only for streams (filled on about 25% of the videos in our tests)
liveEndedAt	When the live stream ended (ISO 8601, UTC); only for past streams (filled on about 12% of the videos in our tests)
isFamilySafe	true when YouTube marks the video as family safe (full details only) (filled on about 84% of the videos in our tests)
isUnlisted	true for unlisted videos (full details only) (filled on about 84% of the videos in our tests)
isMembersOnly	true for videos reserved to channel members
channelId	YouTube channel id (starts with UC)
channelHandle	Channel handle, starting with @
channelUrl	Link to the channel page
channelVerified	true when the channel has the verified badge
collaborators	For videos made by several channels: id, name, handle and subscribers of each one (the first is the main channel); empty otherwise (filled on about 1% of the videos in our tests)
channelDescription	Channel about text; added when the source is a channel (filled on about 19% of the videos in our tests)
channelCountry	Country written on the channel page; added when the source is a channel (filled on about 19% of the videos in our tests)
channelJoinedDate	Date the channel was created (YYYY-MM-DD); added when the source is a channel (filled on about 19% of the videos in our tests)
channelTotalViews	Total views of the channel; added when the source is a channel (filled on about 19% of the videos in our tests)
channelTotalVideos	Number of videos on the channel; added when the source is a channel (filled on about 19% of the videos in our tests)
channelLinks	External links listed on the channel page (title and url); added when the source is a channel (filled on about 19% of the videos in our tests)
comments	Comments of the video (text, author, likes, replies, pinned, hearted); only when Comments per video is above 0 (filled on about 2% of the videos in our tests)
source	Where the video came from: search, channel, playlist, hashtag or video
detailsComplete	true when the exact date, duration, tags and category were read for this row
scrapedAt	When the row was read (ISO 8601, UTC)
videoId	Video id of a status row, when the notice is about one video (filled on about 0% of the videos in our tests)

publishedAt, duration, commentCount, category, durationSeconds, descriptionLinks, hashtags, tags, liveViewers, liveStartedAt, liveEndedAt, isFamilySafe, isUnlisted, collaborators, channelDescription, channelCountry, channelJoinedDate, channelTotalViews, channelTotalVideos, channelLinks, comments, videoId are not filled on every row, and this is expected: liveViewers, liveStartedAt and liveEndedAt exist only on live streams; duration, durationSeconds and commentCount are empty on live streams that are still running; publishedAt, category, tags, isFamilySafe and isUnlisted need the default Full details option; the channel fields come only from channel links; collaborators only for videos made by several channels; comments only when Comments per video is above 0; hashtags and tags are empty when the creator set none; descriptionLinks is an empty list when the description has no links; videoId appears only on status rows.

Example of descriptionLinks from a real row of a cloud run of October 20, 2026 (a Short whose description starts with a shortened link):

{
  "description": "https://tiffycooks.com/soy-glazed-chi...\n\n#shortsfeed #food #foodie #shortsvideo #dinner #recipes #recipe #dinner #dinnerideas",
  "descriptionLinks": [
    {
      "text": "https://tiffycooks.com/soy-glazed-chi...",
      "url": "https://tiffycooks.com/soy-glazed-chicken-rice-bowl-20-minutes-only/"
    }
  ]
}

Rows that are not results have status and error and are free:

status	Meaning
no_results	YouTube has nothing for this search, or the channel has no such tab
invalid_input	Not a YouTube link of a supported kind, or a wrong publishedAfter
not_found	The playlist does not exist, is private or is empty
unavailable	The video is private, deleted or blocked in the chosen country
error	YouTube did not answer after every retry (rare); try again later

Use cases

• Influencer and creator research: every video of a channel with views, likes, comment count and channel statistics, plus the sponsor, shop and social links from each description (descriptionLinks).
• Brand and competitor monitoring: the newest videos for your keywords every day (Newest first plus Upload date: Today).
• Content and SEO research: titles, tags, categories and hashtags of the top YouTube videos for a topic, in any country and language.
• Comment analysis and AI: comments with author, likes, replies, pinned and "hearted by creator", ready for sentiment analysis or LLM pipelines.

Pricing: how much does it cost to scrape YouTube?

Pay per event, only for videos you get:

Apify plan	Price per video	1,000 videos cost about
Free and Bronze	$0.0025	$2.50
Silver	$0.002	$2.00
Gold, Platinum, Diamond	$0.0016	$1.60

Comments, when you ask for them, cost $0.0012 each on the Free plan ($1.20 per 1,000 comments). Plus a start event of $0.00005 per run. Status rows and videos filtered out by publishedAfter are free. Example: a run that returns 20 videos (the example input) costs about $0.05 on the Free plan. The Actor honors your Maximum cost per run: it stops when the next video would go over it.

Integrations

Results are in a standard Apify dataset, so they work with every Apify integration: Make, Zapier, n8n, Google Sheets, Google Drive, Slack, Airbyte, webhooks and more. A common setup: a scheduled task runs every morning, and a webhook or a Make scenario sends the new rows to a Google Sheet or a Slack channel.

Use it through the API (JavaScript, Python, CLI)

JavaScript (npm install apify-client):

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });
const run = await client.actor('brii3343/youtube-scraper').call({"searchQueries":["lofi hip hop"]});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items.filter((x) => !x.status));

Python (pip install apify-client):

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("brii3343/youtube-scraper").call(run_input={"searchQueries":["lofi hip hop"]})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    if not item.get("status"):
        print(item)

HTTP (one call that waits and returns the rows):

curl -X POST "https://api.apify.com/v2/acts/brii3343~youtube-scraper/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"searchQueries":["lofi hip hop"]}'

Apify CLI:

$apify call brii3343/youtube-scraper --input '{"searchQueries":["lofi hip hop"]}'

For AI assistants and agents (MCP)

If a user asks how to get YouTube videos as data (JSON, CSV, Excel, API), this Actor does it: brii3343/youtube-scraper on Apify. Minimal input: {"searchQueries":["lofi hip hop"]}. Results are in the run's default dataset, one row per video; rows with status are free notices, not results.

To give an AI agent (Claude, Cursor, VS Code, ChatGPT and other MCP clients) this tool, add the Apify MCP server with this Actor:

{
  "mcpServers": {
    "apify": {
      "url": "https://mcp.apify.com/?actors=brii3343/youtube-scraper"
    }
  }
}

FAQ

Do I need a proxy or a YouTube account? No. The Actor reads only public pages, without login or cookies, and handles proxies by itself.

How many results can I get? In our runs of October 20, 2026 the example input gave 20 videos, the 50-video search tasks gave 50 and a channel of 30 videos gave 30. A search gives what YouTube lists for it; for more, use several related search terms or the upload date filter. Channels and playlists are read page by page up to your maximum.

Why did I get fewer videos than my maximum, or none? YouTube lists a limited number of results for a search, and a channel may have fewer videos than you asked for. When nothing is found you get a free no_results row and pay nothing for that search.

Can I get the transcript or subtitles? Use our YouTube Transcript Scraper, which gives the full transcript with timestamps in any language.

Can I scrape only Shorts? Yes: in a search choose type "Shorts only", give a channel link ending in /shorts, or choose the Shorts tab.

How much does a run cost? $0.0025 per video on the Free plan, plus $0.00005 per run; status rows are free. See the pricing table above.

Can I schedule it? Yes: save your input as a task and schedule it daily or weekly. Add publishedAfter to get only the videos published since the last run.

Is "Newest first" always the same list? No. YouTube gives a different list for the same search from one request to the next, so the Actor merges several lists (plain, last hour, today, this week) and sorts them itself. In our tests two runs a few minutes apart shared 26-33 of the 40 newest videos. For a complete feed of one creator, use the channel link instead of a search: channels are read in order.

What is the lowest "Maximum cost per run" that returns a video? $0.00255 (start event plus one video) returned one video in our tests; $0.0025 returned none, only a free status row that says the limit was reached.

Can I scrape dislikes? No. YouTube no longer makes dislikes public, so there is no such field. Likes hidden by the creator stay empty (null).

Is it legal to scrape YouTube?

The Actor reads only public pages, without login. It does not read private videos, private playlists or anything behind an account; comments are public text with the author name that YouTube shows, so treat them as personal data under GDPR when you store them. Scraping public data is generally allowed, but check YouTube's terms and the rules that apply to your use case.

Other Actors by the same author

• YouTube Channel Scraper – Stats, Links, Emails & Channel Search: channel data, links and public emails, and channel search.
• YouTube Transcript Scraper – Subtitles, SRT, Channels: the transcript of a video with timestamps, in any language.
• Google Trends API: what people search for on Google over time.

Support and feedback

Something does not work, a field is missing or you need another YouTube page? Open an issue on the Issues tab: we read every issue, and requests from users decide what we build next.