YouTube Scraper
lantfly/youtube-scraper
Videos and Shorts from YouTube searches, hashtags and channels, without the Data API quota.
Run this actor in the dashboard
809 runs quality score 70.15/100 social-media videos updated 2026-11-16
Price
Pay per event, on top of the compute the run uses.
| Event | What it is | USD each |
|---|---|---|
result_itemResult item |
0.002 |
Input
| Field | Type | Default | What it is |
|---|---|---|---|
channelsChannels |
array | [] | @handle or channel URL. |
hashtagsHashtags |
array | [] | Without the #. |
postUrlsPost URLs |
array | [] | Watch, Shorts or youtu.be URLs: one item each from the video page with the exact date, views, likes, comments and subscribers. |
maxResultsResults per search/hashtag/channel |
integer | 30 | |
strategiesStrategies |
array | Optional: which ways to try, in order (names or kinds; `*` = the rest in the default order). Default: http, ytdlp, browser, then official-api/vendor sources when configured. Example: ["ytdlp", "*"]. | |
commentUrlsComment URLs |
array | [] | Post URLs whose comments to read: one target `comments:<url>` each, rows with `kind: "comment"` in the `streamers~youtube-comments-scraper` shape. |
searchQueriesSearch terms |
array | [] | Newest uploads first. |
commentsPerPostComments per post |
integer | 50 | Most comments read per comment URL (top-level comments, newest first where the platform allows). |
maxRequestRetriesMax request retries |
integer | 3 | Retries per target on a new exit IP after a block, rate limit or network error. |
proxyConfigurationProxy |
object | {"groups":["DATACENTER"],"useLantflyProxy":true} |
Sample input
{
"maxResults": 20,
"searchQueries": [
"velyniq"
]
}
README
YouTube Scraper
Comments: commentUrls + commentsPerPost give rows with kind: comment in the streamers~youtube-comments-scraper shape (cid, comment, author, voteCount, replyCount, date, videoId, pageUrl): Data API commentThreads.list with a key, else the watch page's InnerTube next continuation, else yt-dlp --write-comments, else Apify.
Reads YouTube's own web pages (search sorted by upload date, hashtag pages, channel video tabs) and pages through them like the site does. Output follows streamers~youtube-scraper: id, title, url, type (video|short), channelName, channelUrl, channelHandle, thumbnailUrl, viewCount, duration, date (from the relative time, approximate), text (snippet), hashtags.
Likes and comment counts are not on listing pages and are left out. postUrls read each video page instead and add likes, commentsCount, numberOfSubscribers, channelId, channelAvatarUrl, inputUrl, with date exact (publishDate) and text the full description.
Datacenter proxies are enough most of the time; a consent page is accepted with a cookie, a bot check retries on a new exit IP, and after every retry the target is blocked.
Issues
0 open issues0 closed issues
What people reported and what the author answered is not published: sign in to read the reports or open one.
Reviews
No published review yet.