LinkedIn Ad Library Scraper

Scrape public LinkedIn ads by company, advertiser name, keyword, or ad URL. Get ad copy, images, video links, carousel cards, and the details LinkedIn shows about the payer, dates, impressions, and targeting.

Use the results for competitor ad research, creative reviews, campaign research, or AI tools. No LinkedIn account or cookies needed. Export to JSON, CSV, or Excel, or use the Apify API in your own workflow.

What you get

• Exact company searches. Use a company ID or company URL. The Actor checks the advertiser before returning a match.
• Two price options. Choose search cards for a lower cost, or full mode for available ad details.
• Several searches in one run. Each unique ad is saved once. A separate report shows which searches found it.
• Creative links and ad copy. Get images, video files, ordered carousel cards, and available document previews.
• Clear missing data and search limits. See which fields LinkedIn did not show and whether a run reached its result, time, or request limit.

Full detail extraction has been tested live for single image, video, carousel, document, and thought leader ads. Other formats may return search card data with a warning. Document results include available preview data, which can be fewer pages than the source lists. The original PDF is not guaranteed.

Quick start

1. Click Try for free and open the Actor input.
2. Add a company name, company URL, company ID, keyword, or direct ad URL.
3. Choose listing or full and set your result limits.
4. Start the run. Open the dataset to view or download the ads, and check the run summary for search coverage.

For an exact company search, paste this into the JSON input editor:

{
  "companyIds": ["68529"],
  "mode": "full",
  "maxResults": 25,
  "maxResultsPerSearch": 25
}

68529 is HubSpot's LinkedIn company ID. Keep all IDs as strings.

Search several companies or keywords

Each entry is a separate search. The global result limit counts unique ads across all searches.

{
  "searches": [
    {
      "type": "companyUrl",
      "value": "https://www.linkedin.com/company/hubspot/",
      "label": "HubSpot exact company"
    },
    {
      "type": "advertiser",
      "value": "Apify"
    },
    {
      "type": "keyword",
      "value": "marketing automation",
      "filters": { "countries": ["DE"] }
    }
  ],
  "mode": "listing",
  "maxResults": 100,
  "maxResultsPerSearch": 50
}

Get specific ads by URL

{
  "adUrls": [
    "https://www.linkedin.com/ad-library/detail/1578348084",
    "https://www.linkedin.com/ad-library/detail/1508101043"
  ],
  "mode": "full",
  "maxResults": 2,
  "allowListingFallback": false
}

These public URLs were tested during development. LinkedIn can change their availability. Direct ad URLs always request the detail page, even if mode is listing. Search filters do not apply to direct ad URLs.

Listing or full mode?

	listing	full
Best for	Finding ads and reviewing search cards	Reviewing available ad copy, creative details, and ad disclosures
Ad copy	Search card copy; may be shortened	Available detail-page copy; accepted full results are not shortened
Creative assets	Links shown on search cards	Available images, video variants, carousel cards, and document preview details
Payer, dates, impression ranges, targeting	Not requested from detail pages	Returned when LinkedIn shows them
Result price	Listing price	Full price for full results; listing price for a fallback

full is the default. If a detail page fails or uses an unverified format, the default is to save useful search card data at the listing price. These rows have retrievalLevel: "listing" and a warning. Set allowListingFallback: false to skip them instead.

Exact company searches may need detail requests to check the advertiser even in listing mode. Direct URL results have no separate search card to fall back to.

Choose the right search

Input	Meaning
companyIds	Exact LinkedIn organization IDs, supplied as strings
companies	Advertiser names or public LinkedIn company URLs
keywords	Text searches across ad copy
adUrls	Specific public ad detail URLs
searches	Searches with a type, value, optional label, and optional filters
startUrls	Public LinkedIn Ad Library search or detail URLs, or company URLs

Supported searches[].type values are companyId, companyUrl, advertiser, keyword, searchUrl, and adUrl.

A company name is a text search. It can match more than one advertiser. Use a company ID or company URL when you need an exact match. A company URL must resolve to a verified organization ID before that search can return exact matches.

You can combine the input arrays. Invalid individual searches appear in RUN_SUMMARY.inputErrors; other valid searches continue. Invalid global settings stop the run.

Filters and limits

Use filters for all searches, or add filters to an individual searches entry to override them.

{
  "companies": ["HubSpot"],
  "mode": "full",
  "filters": {
    "countries": ["DE", "FR"],
    "payer": "HubSpot",
    "dateRange": "custom",
    "startDate": "2026-09-01",
    "endDate": "2026-09-30"
  },
  "maxResults": 25
}

• countries: Two-letter country codes. This is LinkedIn's ad delivery filter, not the advertiser's headquarters.
• payer: Text search for the company or person paying for the ad.
• dateRange: all-time, last-30-days, current-month, current-year, last-year, or custom.
• startDate and endDate: Both required for custom dates, in YYYY-MM-DD form.
• adFormats: An optional filter applied after ads are read. Values include SINGLE_IMAGE, VIDEO, CAROUSEL, DOCUMENT, TEXT, MESSAGE, CONVERSATION, THOUGHT_LEADER, EVENT, SPOTLIGHT, and EMPLOYER_BRAND. A filter value does not mean that format has verified full detail support.

For searchUrl input, the URL's filters apply. Conflicting structured filters are rejected. Unsupported filters are also rejected.

Setting	Default	What it controls
maxResults	100	Maximum unique ads saved across the run
maxResultsPerSearch	100	Maximum unique matches for each search
maxRequests	5,000	HTTP attempts, including retries and company lookup
maxRunSeconds	900	Active run time in seconds
maxConcurrency	8	Maximum requests at once; allowed range is 1 to 20
allowListingFallback	true	Save useful search cards when full details cannot be accepted
includeEvidence	false	Add a short source-field excerpt and hash for checking extraction

Set either result limit to 0 to remove that row limit. Time, request, and spending limits still apply. resultsLimit is an alias for the global maxResults. skipDetails: true maps to listing; false maps to full. Conflicting aliases are rejected.

Apify cloud runs use datacenter proxy by default. proxyConfiguration can override this. LinkedIn can block requests or slow them down, so run times and available results can vary.

Output fields

The dataset has one row per unique ad within a run. Uniqueness uses both sourceKind and adId. Repeated matches across searches do not create extra rows.

Fields	What they contain
adId, sourceKind, adUrl	Ad ID, library type, and source ad link
advertiserName, advertiserCompanyId, advertiserUrl, advertiserLogo	Advertiser details
contentAuthor, paidBy	Source-shown author and payer, separate from the advertiser
format, formatRaw	Standard format label and LinkedIn's original label
headline, body, bodyTruncated, ctas	Ad text, whether the body is shortened, and button labels
imageUrl, videoUrl, documentUrl, media	Available creative links and media details
carouselCards, document, creativeVariants	Ordered cards, document preview details, and available variants
clickUrl, landingPageUrl	Destination links when the source shows them
availability	Source-shown firstShown and lastShown dates
impressions, impressionsPerCountry	Impression range and country shares, with original labels and parsed values
targeting	Source-shown audience settings, including included or excluded groups
retrievalLevel, fieldStates, warnings	Detail level, missing-field explanations, and issues to review
sourceUrl, scrapedAt, schemaVersion	Source link, collection time, and output version

IDs remain strings. Video results keep available MP4 variants and use the highest bitrate as the main videoUrl. Media links can expire. Asset files are not downloaded.

Missing fields and data accuracy

Full mode means available detail fields passed the Actor's checks. It does not mean LinkedIn showed every field.

Missing values stay null or empty. fieldStates explains why:

State	Meaning
available	The source showed the field
not_exposed	The page was read, but did not show the field
not_requested	This mode did not request the field
restricted	The source reported a restriction
fetch_failed	The source request failed
parse_failed	The response could not be read reliably
unknown	The field's status could not be established

The Actor does not estimate ad spend, turn impression ranges into exact counts, or guess whether an ad is active from its last shown date. activityStatus remains unknown unless LinkedIn explicitly reports a restriction.

For documents, document.previewComplete and document.originalDocumentComplete show the limits of the returned assets. Available cover images and a source manifest link can be returned without a complete preview or the original PDF. The manifest is not fetched.

Check search coverage

A successful Apify run can contain useful partial results. Check these key-value store records before treating the dataset as a complete collection:

Record	What to check
RUN_SUMMARY	Counts, outcome, inventoryComplete, detailsComplete, limits, errors, and billing
QUERY_RESULTS	Matches, page count, exact company resolution, completeness, and stop reason for each search
QUERY_MATCHES	Which searches matched each saved ad
RESOLUTION_RESULTS	How company URLs were resolved to IDs
HTTP_METRICS	Request attempts, retries, blocks, and HTML bytes received

inventoryComplete: true means the source confirmed the end of the requested search. Reaching maxResults means partial coverage, even when the run succeeded. For direct ad URLs, completeness refers to those URLs, not all ads from the advertiser. detailsComplete is a separate check on requested full details.

A confirmed no-match search returns zero rows and an empty outcome. A blocked or unreadable response is reported as an error, never as a confirmed empty search.

API and AI agent use

Use the same JSON input through the Apify API. This example starts a small full-mode run and waits up to 60 seconds:

curl --request POST \
  'https://api.apify.com/v2/acts/whoareyouanas~linkedin-ad-library-scraper/runs?waitForFinish=60' \
  --header "Authorization: Bearer $APIFY_TOKEN" \
  --header 'Content-Type: application/json' \
  --data '{"companyIds":["68529"],"mode":"full","maxResults":5,"maxResultsPerSearch":5,"maxRunSeconds":90}'

The response includes data.id, data.status, data.defaultDatasetId, and data.defaultKeyValueStoreId. The run may still be running after the wait. Check GET /v2/actor-runs/{runId} until it finishes, then read:

GET https://api.apify.com/v2/datasets/{defaultDatasetId}/items?format=json
GET https://api.apify.com/v2/key-value-stores/{defaultKeyValueStoreId}/records/RUN_SUMMARY
GET https://api.apify.com/v2/key-value-stores/{defaultKeyValueStoreId}/records/QUERY_RESULTS

Send the same authorization header with these requests. Use format=csv for a CSV export.

For agents: Keep IDs as strings. Use exact company IDs or company URLs for exact advertiser research. Set result and run limits. After each run, read RUN_SUMMARY, QUERY_RESULTS, retrievalLevel, and fieldStates. Treat missing fields as unknown data, and partial searches as partial coverage. Do not use empty arrays or old dates as proof that a company has no active ads.

Pricing

Pay for unique ads saved, plus a small run-start charge. Standard in-run computing and the default proxy are included. Standard Apify storage charges can apply after the run.

Apify price tier	Listing, per 1,000 ads	Full, per 1,000 ads
Free	$0.40	$1.20
Bronze	$0.35	$1.00
Silver	$0.30	$0.85
Gold, Platinum, Diamond	$0.25	$0.70

The start charge is $0.002 per memory-based start unit. At the default 512 MB, a run uses one start unit, so the start charge is $0.002. At up to 1 GB, it is still one unit. Above 1 GB, Apify rounds the assigned memory up to whole GB and charges that many units.

Billing events are ad-listing, ad-full, and apify-actor-start. A saved ad is charged at one result level only. Full-mode rows that fall back to listing data use the listing price.

• Duplicate matches within the same run have no extra ad charge.
• Failed or skipped ads have no ad charge.
• Empty and failed runs still have the run-start charge.
• Each new run is billed separately, including ads saved in an earlier run.

For example, 1,000 accepted full results on the Free plan at the default memory cost $1.202, including one start unit. Check the Actor's Pricing tab for the current rates and set a spending limit before starting a large run. Custom proxies and services you supply can have their own costs.

FAQ

Do I need a LinkedIn account?

No. The Actor reads the public LinkedIn Ad Library. It does not need your login or cookies.

Does it scrape LinkedIn profiles or contact details?

This Actor collects public ad library records. It does not collect member profiles, emails, or private LinkedIn data.

Can I get all ads from one company?

Use an exact company ID or company URL and choose suitable limits. The Actor follows search pages, but access blocks and budgets can stop the run. Check inventoryComplete and the per-search stop reason before calling the collection complete.

Are targeting and impression details always available?

No. Availability depends on what LinkedIn shows for that ad and region. Use full mode, then check fieldStates to distinguish missing source data from a request failure.

Can I download the original document or video?

The Actor returns available source links. Video ads can include direct MP4 links. Document ads may expose only covers and preview details. An original PDF is not promised, and files are not downloaded by the Actor.

Which formats have full detail support?

Single image, video, carousel, document, and thought leader layouts have passed live full-detail checks. Other formats can return listing data with warnings. Set allowListingFallback: false if you want accepted full results only.

How fast is it?

Speed depends on the number of ads, mode, and LinkedIn response times. Tested cloud runs returned 50 full ads in about 53 seconds and three direct video, carousel, and document ads in about 3 seconds. These are examples, not a speed guarantee.

Can I run it on a schedule?

Yes. Use Apify schedules to start new runs. Each run returns its own dataset; this Actor does not compare runs or create change alerts.

Support

Open an issue on this Actor's Issues tab with the run ID and a description of the problem. Include the search type and expected result so the issue can be reproduced. Do not share account tokens or other secrets.