A YouTube channel scraper API lists a channel's videos with their metrics — title, video id, views as text and as a number, publish age, duration and thumbnail — and stamps every row with the channel's own name, id, description, subscriber count and total video count.
That stamping is not padding. It means a single flat table answers both the channel-level question (how big is this creator) and the video-level question (how does each upload perform), with no join and no second call. Point it at a handle, a channel id or a channel URL, choose the videos, shorts or streams tab, and it reads the page's own data payload — no Data API key, no quota, no browser.
$0.0008 per delivered video, up to 60 videos per run. Nothing delivered means nothing charged.
The channel input accepts a handle (@NASA), a channel id or a full channel URL, and resolves all three. The tab input picks which grid to read: regular uploads, shorts, or past live streams. Those are genuinely different catalogues on modern YouTube, and treating them as one — which most scrapers do — produces a nonsense view of a channel that posts daily shorts and monthly long-form.
One page per run, roughly 30 videos. YouTube puts the rest of the grid behind a continuation token, and this collector does not pretend to walk it. For most work the most recent 30 uploads is the question anyway; for a full back-catalogue you would want a different approach and a different budget.
live flags anything currently streaming, and duration is null for it, because a live stream has no final duration and a running view counter is not a comparable metric.
channel is required and accepts a handle, an id or a URL. tab selects uploads, shorts or streams — pick deliberately, since a channel's shorts and its long-form videos have completely different view distributions and averaging them together is meaningless. country sets the exit region; max_results caps delivery within the single page a run covers.
One row per video, with channel-level fields repeated on each. channel_subscribers and channel_videos are the channel's own totals, so engagement relative to audience size needs no second lookup. views_value is the integer behind the views display string. published is YouTube's relative age rather than a timestamp — that is what the grid publishes, so treat it as approximate rather than as a date. channel_description is the About blurb, which is often where a creator states their contact and business details.
$0.0008 per delivered video ($0.80 per 1,000). Nothing delivered means nothing charged, and the $2 monthly free credit covers roughly 2,500 videos here. Volume tiers take up to 30% off.
$0.0008 per delivered video, 80 cents per 1,000. Tracking 200 channels weekly at 30 videos each is 6,000 rows, about $4.80 a week. A one-off competitive audit of 50 channels is well under a dollar.
Blocked pages and retries are never billed, and the free $2 monthly allowance covers about 2,500 videos.
For a channel's uploads specifically, the Data API is cheaper on quota than search: you fetch the uploads playlist and page through it at 1 unit per call rather than 100. If you are reading a handful of channels, the official API is a perfectly good answer and it is free.
Where it stops being convenient is breadth and shape. Getting from a handle to an uploads playlist to video statistics is three different calls across three resources, and view counts require a separate videos.list pass because the playlist items do not carry them. Multiply by a few hundred channels and the quota arithmetic returns.
The shorts and streams tabs are the other gap: the Data API does not model them as distinct catalogues, so separating a channel's shorts from its long-form output means inferring it from duration and hoping. Here it is an input.
The recurring problems in home-grown channel scrapers are specific:
@handle, legacy /c/ URLs, /user/ URLs and raw channel ids all coexist, and each resolves differently.For search rather than a specific channel, the YouTube scraper API covers that side.
Weekly runs across a competitor set give publishing cadence, format mix and per-video performance in one table.
Subscriber count is easy to inflate; median views across the last 30 uploads is not. Both arrive in the same rows.
Run the same channel with tab set to videos and then to shorts to see which format actually carries the audience.
One scheduled job across every managed channel, without connecting to anyone's account or asking for OAuth.
One grid page per run, roughly 30 videos, up to 60. The rest of a channel's back-catalogue sits behind continuation tokens and is out of scope. More channels parallelise up to your plan's rate limit — 60 requests/minute on pay-as-you-go, up to 1,200 on the top tier.
Legally: channel and video metadata are publicly displayed, and collecting public data is generally lawful in most jurisdictions, but YouTube's terms restrict automated access. Only metadata and links are returned, never media. Creator names and channel descriptions can be personal data. Not legal advice.
No. Your QuantumProxies key is the only credential — no Google Cloud project, no OAuth, no quota. The Data API can also list a channel's uploads cheaply on quota, but it takes three calls across three resources to get videos with view counts, and it does not model the shorts and streams tabs separately.
There is a free allowance of $2 a month with no card, about 2,500 delivered videos at $0.0008 each — enough to audit a competitive set of 80 channels before spending anything.
One grid page, roughly 30 videos and up to 60. YouTube puts the remainder of a channel's catalogue behind a continuation token, which this collector does not walk — the most recent uploads are what it covers.
Yes, and you should. The tab input selects videos, shorts or streams. They are genuinely different catalogues with different view distributions, so averaging them together misrepresents any channel that posts both.
$0.0008 per delivered video — 80 cents per 1,000 — with no subscription. Volume tiers take up to 30% off, and blocked pages are never billed.
Yes. channel_subscribers and channel_videos are stamped on every row alongside the channel name, id and description, so you can compute per-video performance relative to audience size from a single flat table with no join.
A handle like @NASA, a raw channel id, or a full channel URL. All three are resolved, so you can feed whatever form your existing data holds without normalising it first.
Skip the marketing. QuantumProxies.io publishes a machine-readable site map, Markdown for every page, a free MCP server, and APIs billed from the same balance as your proxies.
Connect in one command: npx -y quantumproxies-mcp