YouTube Shorts Sponsorship Signals Scraper

---

🎯 What this scrapes

Influencer-intel SaaS dashboards charge $200–2 000/mo to tell agencies which creators run paid placements. This Actor surfaces the same signal from public data for a fraction of the cost. For every recent Short on a channel it pulls the title, view count, duration, tags, description, and — crucially — the spoken transcript, then scores how likely the Short is sponsored and lists the exact phrases that fired.

Shorts descriptions are mostly empty, so caption-only scrapers miss most deals. We read what the creator actually says, which is where Shorts sponsorships live.

🔥 What we handle for you

• 🛡️ Browser fingerprint impersonation — curl-cffi mimics a real Chrome TLS + HTTP/2 handshake, so YouTube sees a browser, not a bot.
• 🍪 Consent-gate handling — we send the consent cookie YouTube requires before it will list a channel's Shorts. Skip it and you get zero results; we don't skip it.
• 🗣️ Transcript-first detection — spoken sponsor language is the primary signal, caught even when the description is blank.
• 🌐 Optional proxy rotation via Apify Proxy — datacenter is fine at low volume; residential scales it up.
• 🧱 Graceful degradation — a Short with no captions still emits a row (has_transcript: false) instead of crashing the run.
• 💰 Pay-Per-Event pricing — you only pay for Shorts that land in your dataset.

💡 Use cases

• Influencer-agency intel — see which creators run brand deals and which brands keep coming back.
• Competitive brand monitoring — track where a competitor's products show up across creators' Shorts.
• Sponsorship benchmarking — measure how often a niche's top channels run paid placements.
• Disclosure compliance audits — flag Shorts with spoken sponsorships but no #ad disclosure.
• Creator vetting — gauge a channel's commercial density before signing a deal.

⚙️ How to use it

1. Click Try for free at the top of the page.
2. Add one or more channel handles — @mkbhd, mkbhd, or a full youtube.com/@mkbhd URL all work.
3. Click Start. Rows stream into the run's dataset as each Short is scored.
4. Export from Storage → Dataset as JSON, CSV, or Excel — or fetch via the API.

📥 Input

Field	Type	Required	Default	Notes
channels	array	yes	["@mkbhd"]	Channel handles or URLs. Normalised to @handle.
maxShortsPerChannel	integer	no	20	1–200. How many recent Shorts to inspect per channel.
language	string	no	"en"	Preferred transcript language; falls back to any available track.
minSponsorshipScore	number	no	0.0	0–1. Only emit rows scoring ≥ this. Set 0.5 to keep only likely-sponsored Shorts.
proxyConfiguration	object	no	{"useApifyProxy": false}	Optional. Direct (no proxy) by default — YouTube degrades the player and transcript endpoints for shared datacenter IPs. Enable residential proxy only if you hit rate limits at high volume.

Example input

{
  "channels": ["@mkbhd"],
  "maxShortsPerChannel": 20,
  "language": "en",
  "minSponsorshipScore": 0.0,
  "proxyConfiguration": { "useApifyProxy": false }
}

📤 Output

Every row is one Short.

Field	Type	Notes
channel	string	Channel handle (@handle).
video_id	string	YouTube Short ID (11 chars).
url	string	Canonical Shorts URL.
title	string	Short title.
published_text	['string','null']	Human-readable publish text when available.
view_count	['integer','null']	Views at scrape time.
length_seconds	['integer','null']	Duration in seconds.
description	string	Short description (often empty for Shorts).
tags	array	Video keywords/tags.
has_transcript	boolean	True when a spoken transcript was retrieved.
transcript_chars	integer	Character length of the transcript.
hashtags	array	#hashtags found in description/transcript.
mentions	array	@mentions found in the description.
detected_brands	array	Brands inferred from mentions, domains, and tokens after code/sponsored by.
sponsorship_signals	array	The literal phrases/tokens that fired.
sponsorship_score	number	Heuristic sponsorship likelihood, 0.0–1.0.
is_likely_sponsored	boolean	True when sponsorship_score >= 0.5.
scraped_at	string	ISO-8601 timestamp of the scrape.

Example output

{
  "channel": "@mkbhd",
  "video_id": "n3V3LZh_r40",
  "url": "https://www.youtube.com/shorts/n3V3LZh_r40",
  "title": "My everyday carry, sponsored",
  "published_text": null,
  "view_count": 2805352,
  "length_seconds": 58,
  "description": "#ad",
  "tags": ["tech", "edc"],
  "has_transcript": true,
  "transcript_chars": 712,
  "hashtags": ["#ad"],
  "mentions": [],
  "detected_brands": ["Surfshark", "MARQUES"],
  "sponsorship_signals": ["this video is sponsored", "use code", "#ad"],
  "sponsorship_score": 1.0,
  "is_likely_sponsored": true,
  "scraped_at": "2026-06-13T10:00:00+00:00"
}

🧮 How the score works

A transparent weighted sum over transcript + description + title, clamped to [0, 1]:

Signal	Examples	Weight
Spoken strong	sponsored by, this video is sponsored, for sponsoring, use code, promo code, paid partnership	+0.5 each
Spoken weak	check out, link in bio, link in description, discount, % off	+0.15 each
Disclosure hashtag	#ad, #sponsored, #spon, #partner, #partnership, #collab	+0.4 each
@mention / brand domain	@brand, brand.com (in description)	+0.1 each

A single strong spoken signal crosses the 0.5 threshold and marks the Short as likely sponsored. The fired phrases are returned in sponsorship_signals so you can audit every call.

💰 Pricing

Pay-Per-Event — you pay only when these events fire:

Event	USD	What it is
actor-start	$0.005	One-off warm-up charge per run
result	$0.004	Per Short written to the dataset

Example: 1 000 Shorts ≈ $4.00. No subscription, no minimum, no card to start — Apify gives every new account $5 of free credit.

🚧 Limitations

The score is a heuristic, not a legal determination — treat is_likely_sponsored as a strong lead, not proof. Some Shorts (music-only, no narration) have no captions; those still emit a row with has_transcript: false and score only on description/hashtag signals. YouTube's official "Includes paid promotion" flag is not third-party accessible, so we infer from spoken + textual signals instead.

❓ FAQ

Why transcripts instead of captions/descriptions?

Because Shorts descriptions are almost always empty. The sponsorship lives in what the creator says, so we read the transcript first and treat description/hashtags as secondary signals.

What handle formats are accepted?

@mkbhd, mkbhd, youtube.com/@mkbhd, and full https://…/@mkbhd/shorts URLs all normalise to @mkbhd.

Do I need a proxy or API key?

No. The data path is keyless and works without a proxy at low volume. Add a proxy if you're scanning many channels.

Can I keep only the sponsored Shorts?

Yes — set minSponsorshipScore to 0.5 and only likely-sponsored Shorts are emitted (and billed).

💬 Your feedback

Spotted a bug, hit a weird edge case, or need a new field? Open an issue on the Actor's Issues tab on Apify Console — we ship fixes weekly and we read every report.

---