YouTube Channel Scraper – Videos, Views & Likes
Pricing
from $0.30 / 1,000 videos
Scrape the latest videos from any public YouTube channel — title, link, exact view count, likes, publish date, duration, description and thumbnail — one row per video. No login, no cookies, no API key and no browser: a channel listing costs one request. Unofficial.
Channel addresses, one per entry. All four of YouTube's forms work: https://www.youtube.com/@NASA, /channel/UC…, and the legacy /c/ and /user/ forms. Give a channel, not a video — a watch URL comes back as an error row telling you so.
The same thing as a plain list — easier to paste a column out of a spreadsheet. Accepts "@NASA", bare "NASA", or a channel ID like UCLA_DiR1FfKNvjuUpBHmylQ. Both inputs are read; giving either is enough.
[ "@NASA"]How many of the channel's latest videos to return, newest first. A channel page renders 30 to a logged-out reader and offers no way to page further without YouTube's internal API, so 30 is the ceiling — every row carries channelVideoCount so you can tell when you are seeing all of them.
On by default. A channel listing rounds every view count ("53K views") and gives only a relative date ("3 days ago"). This reads each video's own page for the exact figures — 53,674 views and the real publish date — and adds the like count and description, which appear nowhere else.
It is also what the price turns on. Reading a channel listing is one request for up to thirty videos; the exact figures are a request each, and measured that is five times the cost. So a video costs $0.0005 either way and the exact figures add $0.0010 on top — $1.50 per 1,000 videos with them, $0.50 per 1,000 without. Turn this off for a cheap freshness check across many channels.
Every read that lands carries the exact view count, the like count, the description and the publish day. The publish time of day, the category and the keywords come only when YouTube also sends its player data, which it withholds from most datacenter requests; those rows carry the date as a day and say so via publishedAtIsApproximate. A video whose page could not be read at all keeps the listing's rounded figures, says so via viewCountIsApproximate, and is not charged the detail fee.
Keep only videos published after this point — an ISO date like 2026-08-01, or a window like "7 days", "3 months". If the window reaches further back than the 30 videos a channel page shows, the channel comes back as an error row (window_too_wide) rather than a short list that would look like a complete answer.
Defaults to Apify's datacenter proxy, which is included in every plan and is what this Actor is built and measured on. Channel listings, and each video's exact views and likes, read reliably on it.
Residential costs roughly ten times as much per video and does not improve those figures. What it adds is YouTube's player data on more videos — the publish time of day, the category and the keywords.
{ "useApifyProxy": true}