LinkedIn Ad Library Transcript Scraper — video hooks, CTAs & ad copy

Turn LinkedIn Ad Library ads into structured text — video ads become full transcripts with the first-3-seconds hook, and image ads have their on-image copy read out — plus advertiser/creative metadata, in one JSON row per ad. Give it advertiser names or keywords (the actor searches the Ad Library for you), or chain it straight after any LinkedIn Ad Library scraper run. You are charged only when text is delivered — creatives with nothing readable, blocked pages and failures are never charged. No cookies, no LinkedIn login.

New: image ads are read too, and that changes your bill. About 4 in 5 LinkedIn ads are static image ads, and each one whose copy is readable is now a charged result at the same price as a transcript. A keyword run that used to return a handful of video transcripts now returns roughly ten times as many charged rows. Use maxItems and the run's max total charge to set the size you want — or turn Read the text on image ads off to get the old video-only behaviour exactly as before.

You give it	You get back
Advertiser names (e.g. HubSpot)	transcript — full speech-to-text, any length
…or keywords (e.g. crm software)	hook3s — what's said in the first 3 seconds
	imageText — the headline, body and CTA read off an image ad
…or ad URLs / video URLs	advertiser, headline, adText, payingEntity, language, durationSeconds, timestamped segments
…or an Ad Library scraper run's dataset ID	one JSON row per ad, misses clearly marked and never charged

Output

One JSON row per ad. Delivered rows carry the transcript (or the creative's text) and charged: true; anything that could not be delivered carries a status saying why, and charged: false. Real rows from the live example run below (long text trimmed with … for the page):

status	charged	advertiser	headline	hook3s	transcript	language	durationSeconds
transcribed	true	Sparkon | HubSpot Elite Partner | Growth, AI, CRM Implementations, RevOps	We are Hubspot Elite Partner	If your revenue team is working harder than ever, but closing less, the problem isn't effort, it's the system.	If your revenue team is working harder than ever, but closing less, the problem isn't effort, it's the system. Too many manual tasks, too many disconnected tools…	English	43.86
transcribed	true	Sparkon | HubSpot Elite Partner | Growth, AI, CRM Implementations, RevOps	We are Hubspot Elite Partner	When your CRM becomes too complex, every team pays the price.	When your CRM becomes too complex, every team pays the price. We connect your financial systems, automate workflows, and give every team clear operational visibility…	English	10.68
image_text_extracted	true	HubSpot	null	null	imageText: "46% of marketers struggle with visual consistency. Get the free HubSpot & Canva Playbook"	null	null
no_audio_stream	false	HubSpot	null	null	null	null	null

Full row, as JSON:

{
  "status": "transcribed",
  "charged": true,
  "advertiser": "Sparkon | HubSpot Elite Partner | Growth, AI, CRM Implementations, RevOps",
  "headline": "We are Hubspot Elite Partner",
  "adText": "Is your revenue team working harder, but closing less? ⚙️ …",
  "payingEntity": "Paid for by Triario LLC",
  "adId": "1508450294",
  "detailUrl": "https://www.linkedin.com/ad-library/detail/1508450294",
  "format": "SPONSORED_VIDEO",
  "availability": "Ran from Aug 17, 2026 to Aug 21, 2026",
  "impressions": "10k-20k",
  "impressionsByCountry": [{ "country": "Germany", "share": "70%" }, { "country": "Switzerland", "share": "22%" }],
  "hook3s": "If your revenue team is working harder than ever, but closing less,",
  "transcript": "If your revenue team is working harder than ever, but closing less, the problem isn't effort, it's the system...",
  "language": "English",
  "durationSeconds": 43.9,
  "segments": [{ "start": 0, "end": 7.7, "text": "If your revenue team is working harder than ever, but closing less, the problem isn't effort, it's the system." }],
  "chargeEvents": { "transcript": 1, "surchargeMinutes": 0 }
}

Browse the full live example dataset → — real output from a verified run, including the free uncharged rows.

An image ad's row looks the same, with imageText in place of the transcript:

{
  "status": "image_text_extracted",
  "charged": true,
  "advertiser": "HubSpot",
  "format": "SPONSORED_STATUS_UPDATE",
  "adId": "1478297483",
  "detailUrl": "https://www.linkedin.com/ad-library/detail/1478297483",
  "imageUrl": "https://media.licdn.com/dms/image/v2/…",
  "imageText": {
    "headline": null,
    "body": null,
    "cta": null,
    "displayUrl": null,
    "rawText": "46% of marketers struggle with visual consistency. Get the free HubSpot & Canva Playbook HubSpot | Canva"
  },
  "chargeEvents": { "transcript": 1, "surchargeMinutes": 0 }
}

advertiser always comes from LinkedIn's own metadata, never from reading a logo in the picture.

Every field is always present (explicit null over silent omission). Non-delivered rows carry status + statusReason instead — ocr_no_text_found (a genuinely text-free creative, e.g. a plain stock photo), image_expired (most creative image links never expire, but a small share — 2 of 76 measured — are short-signed and can go stale in a chained dataset; re-run your scraper for fresh URLs, or search here so the link and the picture are read in the same run), image_skipped (you turned image reading off), non_video_skipped (document, carousel, article, message and text-only creatives pass through with their metadata), no_audio_stream (a genuinely silent video), no_speech (music-only audio), no_onscreen_text, not_found, blocked_by_challenge, unavailable_expired, no_videos_found, search_unavailable, no_creative_found, input_error, or failed_* (also listed in the run's ERRORS record). None of them are charged. availability and impressions are filled when LinkedIn shows them (EU-targeted ads) and null otherwise.

For AI agents & LLMs — copy this block

Actor:   steadyfetch/linkedin-ads-transcript-scraper  (Apify Store)
Does:    LinkedIn Ad Library video ads -> full transcript + first-3-seconds hook +
         advertiser metadata. One JSON row per ad. No LinkedIn login or cookies.
Run:     POST https://api.apify.com/v2/acts/steadyfetch~linkedin-ads-transcript-scraper/run-sync-get-dataset-items?token=APIFY_TOKEN
Input    (any ONE of the first four fields is enough):
  {
    "accountOwners": ["HubSpot"],          // advertiser names; the actor finds their video ads
    "keywords": ["crm software"],          // or search the Ad Library by topic
    "videoUrls": ["https://www.linkedin.com/ad-library/detail/1508450294"],
    "datasetId": "<dataset id of any LinkedIn Ad Library scraper run>",
    "includeImageText": true,              // default ON: read image ads' copy (charged like a transcript)
    "includeOnScreenText": false,          // default OFF: read silent video ads' on-screen text
    "includeNonVideo": false,              // optional: free rows for document/carousel/article/text ads
    "maxItems": 1000                       // optional cap — counts video AND image ads
  }
Output   one row per ad: status, charged, advertiser, headline, adText, ctaText, payingEntity,
         adId, detailUrl, format, availability, impressions, impressionsByCountry,
         hook3s, transcript, language, durationSeconds, segments[],
         imageUrl, imageText{headline,body,cta,displayUrl,rawText}, onScreenText{...}
Pricing  pay per event, platform usage INCLUDED, no start fee:
  video-ad-transcript   $0.020 -> $0.008 per delivered transcript OR image-ad text extraction
                        (falls with your Apify plan tier; an image never adds a surcharge)
  long-video-surcharge  $0.005 per started minute past the first 3 minutes
  Charged only on delivery — creatives with no readable text, blocked pages and failures cost $0.
  ~4 in 5 LinkedIn ads are image ads, so a keyword run charges for ~10x more rows than video alone.
  Cap any run with maxTotalChargeUsd (minimum $0.05).
MCP:     reachable through Apify's MCP server; ask for "linkedin ad transcripts".

Why this LinkedIn Ad Library transcript scraper?

• It reads the whole library, not just the 7% that is video. The Ad Library has no format filter. This actor reads the format off each search result first, transcribes the video ads, and reads the on-image copy of the image ads — so a keyword search comes back with the ad copy of ~95% of what it saw instead of ~7%.
• Image ads cost no extra page loads. The creative image URL is already in the free search result, so reading image ads adds zero requests to LinkedIn.
• Charged only on delivery. Text either lands in your dataset or the row is not charged. Text-free stock photo → uncharged row that says so. Silent creative → uncharged row that says so. Blocked page → uncharged row that says so. No "it ran and charged me but returned nothing."
• The run tells you what it saw. The status line names every format the search returned, including the ones this actor does not read — nothing is dropped silently.
• Old datasets still work. LinkedIn ad video URLs are long-lived — unlike other ad platforms' CDN links, they don't expire in hours or days. A dataset you scraped weeks ago transcribes fine today.
• No cookies, no login. Runs entirely on LinkedIn's public Ad Library — nothing to paste, no account at risk.
• Any video length. Long-form B2B creatives transcribe fine — videos past 3 minutes add a small per-started-minute surcharge instead of failing.
• Hook as data. The first-3s hook is the line media buyers study; you get it as its own field, plus full segment timestamps.
• Works with any scraper's output. The dataset input deep-scans your rows for LinkedIn ad video links and ad detail links — no field mapping, no glue code.
• Ad metadata included. Advertiser, headline, ad text, paying entity, run dates, and — where LinkedIn shows them (EU-targeted ads) — estimated impressions with the per-country split, on every row.

This actor may fail when the platform changes things — failed items are never charged.

How to transcribe LinkedIn video ads (no code)

1. Type an advertiser name (e.g. HubSpot) into Advertiser names — or a topic into Keywords.
2. Click Start. The actor searches the Ad Library, picks out the video ads, and transcribes each one.
3. Each ad comes back as one JSON row: a transcript + hook for video ads, the on-image copy for image ads, plus advertiser metadata.
4. Export as JSON/CSV, or read it via API.

Already scraping the Ad Library? Paste your run's dataset ID (or use Apify's Connect Actor integration) and this actor transcribes the video ads in it.

How much does it cost to transcribe LinkedIn ads?

One result = one ad's text payload — a video ad's transcript + hook + segments, or an image ad's on-image copy. Both cost the same event. The first 3 minutes of each video are included; longer videos add a small per-started-minute surcharge, and an image ad never does. No start fee, no subscription, no third-party API key needed.

Job	Approx. cost
50 competitor ads (video or image)	≈ $1.00
500 ads (a serious creative teardown)	≈ $10
Apify free plan ($5 credit)	≈ 250 ad creatives

Sizing a keyword run: roughly 4 in 5 LinkedIn ads are image ads and about 3 in 4 of those carry readable copy, so one page of 24 search results yields on the order of 15 charged rows. Set maxItems to the number of ads you actually want, and cap the run with max total charge — both stop the run honestly with a "what's left" message rather than overspending.

Platform usage (compute + transfer) is included in the event price — the price you see is the whole price. Searching, unreadable creatives, and failed fetches carry no result fee.

Works with — LinkedIn Ad Library scrapers you can chain from

The dataset input reads the values in each row (any dms.licdn.com video link or ad-library/detail/… link), not fixed field names — so it works with any Ad Library scraper, including:

Scraper	Chain by
silva95gustavo/linkedin-ad-library-scraper	dataset ID or pasted rows (videoUrl / detail links)
dz_omar/linkedin-ads-scraper	dataset ID or pasted rows (detailPageUrl, detail.videoUrls)
memo23/linkedin-ads-scraper	dataset ID or pasted rows (adLibraryUrl + format: "VIDEO")
ivanvs/linkedin-ad-library-scraper	dataset ID or pasted rows (url detail links)
automation-lab/linkedin-ad-library-scraper	dataset ID or pasted rows (detailUrl, mediaUrl)

How to chain: run the scraper → copy the run's dataset ID into this actor's Dataset ID field (or use Apify's Connect Actor integration) → Start. Video rows are transcribed; image rows have their on-image copy read (turn Read the text on image ads off to pass them through free instead); rows with no recognizable creative come back as uncharged no_creative_found. One caveat for chained rows: most creative image links never expire, but a small share are short-signed and can go stale in an older dataset — those rows come back as uncharged image_expired, and searching by advertiser/keyword here avoids the problem entirely. You can also skip the other scraper entirely — this actor searches the Ad Library itself.

Use it via API, MCP, and integrations

• API: standard Apify run API — POST .../acts/steadyfetch~linkedin-ads-transcript-scraper/runs with { "accountOwners": ["HubSpot"] }. Python/Node clients work as with any actor.
• MCP: works from Claude, Cursor, and any MCP client through Apify's MCP server — ask for "linkedin ad transcripts" with an advertiser name or dataset ID.
• n8n: a free ready-made template, no community nodes needed — LinkedIn Ad Library transcripts → (one advertiser → hooks & transcripts → spreadsheet-shaped rows, 3-minute setup).
• Make / Zapier: call it as a regular Apify actor step right after your scraper node, or run it standalone with advertiser names.

FAQ

Where do the ads come from? LinkedIn's own Ad Library — the public transparency surface LinkedIn publishes so anyone can see who runs which ads. No login or cookies are used, ever.

Why did some ads come back non_video_skipped? Document, carousel, article, message and text-only creatives are not read — there is no video to transcribe and no single ad image to read. They pass through uncharged with their advertiser/CTA metadata (turn on Include the remaining formats to get one uncharged metadata row for each one found in search).

How accurate is the image-ad text? It is read by a vision model from the ad's own creative, verbatim and untranslated, and delivered as imageText.rawText with the headline / body / CTA split out where the model can tell them apart. Non-English copy works. If nothing readable comes back, the row is ocr_no_text_found and free — a plain product photo with no copy on it is never sold to you as a result.

Can I get the on-screen text of silent video ads? Yes — turn on Read on-screen text on silent video ads. About half of LinkedIn's video ads are silent or music-only, and most of them carry their whole message as on-screen type. With the switch on, those ads' frames are read and delivered as onScreenText, charged like a transcript; silent ads with no real copy (plain b-roll) stay free. With the switch off — the default — a silent ad is never charged, exactly as before.

A search returned no_videos_found? The advertiser or keyword had no video ads and no image ads this actor could read — only document, carousel, article or message creatives, or nothing at all. The row tells you how many creatives were seen and of which formats, so you can widen the query. When you search several advertisers/keywords at once, each one that turns up nothing readable gets its own uncharged row naming it, so a misspelled advertiser never disappears silently. Nothing is charged.

Silent video ads? Some B2B creatives ship with no audio track at all (text-overlay videos). We detect that and return an honest, uncharged no_audio_stream row instead of an empty transcript — or, with Read on-screen text on silent video ads turned on, we read the type off the video's own frames.

Music-only ads? Speech-recognition models hallucinate filler on music. We detect that and return an honest, uncharged no_speech row instead of selling you a fake transcript.

What is blocked_by_challenge / search_unavailable? LinkedIn occasionally serves a verification page instead of an ad page, or the Ad Library search returns nothing for a while. The actor retries, and if it still can't get through you get an uncharged row that says so — re-running a few minutes later usually picks those ads up.

Typos in bare ad IDs? A mistyped numeric ad ID is indistinguishable from a real one, so it surfaces as an honest, uncharged not_found row (LinkedIn answers "can't load this ad") rather than an input error.

Languages? Whisper-class multilingual ASR — German, French, Spanish, Arabic, Japanese, and ~90 more transcribe out of the box; the detected language ships on every row.

Is it legal to transcribe LinkedIn ads? The Ad Library is public by design (ad transparency, required by regulations like the EU DSA). This actor reads only that public surface, logged out. As always, how you use the data is on you.

Is this an official LinkedIn product? No. This is an independent actor from steadyfetch — unofficial, and not affiliated with, endorsed by, or sponsored by LinkedIn Corporation or Microsoft. "LinkedIn" is a trademark of its respective owner, used here only to name the public Ad Library surface this actor reads.

Steadyfetch ad-intelligence suite

One actor per ad surface, one job each, the same contract everywhere: all-inclusive pay-per-event, no start fee, charged only on delivery — and the same row shape, so a workflow built for one platform ports to the next.

Ad creatives → transcripts (video)

Ad surface	Actor	Free n8n template
Facebook & Instagram Ad Library	Facebook Ad Library Transcript Scraper — Video Hooks & CTAs	Facebook ad transcripts → · Competitor ad teardown →
Google Ads Transparency Center	Google Ads Video Transcripts & Hooks (Ads Transparency Center)	Google Ads video transcripts →
TikTok Creative Center Top Ads	TikTok Ad Transcripts & Hooks (Creative Center Top Ads)	TikTok Top Ads transcripts →
LinkedIn Ad Library	this actor	LinkedIn Ad Library transcripts →
Instagram Reels (organic posts)	Instagram Reels to Text · @handle	Instagram Reel transcripts →

Ad creatives → text (image & text ads, read by OCR)

Ad surface	Actor
Google Ads Transparency Center	Google Ads Creative Text & OCR (Transparency Center)

All templates in one place: github.com/steadyfetch/n8n-templates — free, no community nodes needed.

The rest of the steadyfetch shelf — same contract everywhere: all-inclusive pay per event, no start fee, charged only on delivery.

Family	Actors
Ad creative intelligence	Facebook · Google Ads video · TikTok · LinkedIn · Google Ads text & OCR
Trends & keywords	Google Trends · Trends Now · Breakout keywords · Autocomplete keywords · Keyword volume & CPC · Social trends
YouTube transcripts	YouTube videos · YouTube channels
Instagram	Reel transcripts · Profile posts
Jobs	Indeed · Career sites by domain · Glassdoor · Multi-board
Amazon	Products · Search · Bestsellers · Sellers
Any media file	Speech to Text · any link or file

Unlinked names are publishing shortly on the same account — search steadyfetch on Apify Store.

Feedback & support

Found an issue? Open it on the Issues tab — we respond within one business day. Feature requests welcome.