See what a competitor is paying to say. The Facebook Ad Library is the deep lane: find the advertiser, then pull its live and archived creative filtered by country, status, format and date. Google ads by domain sit on the same key.
What rivals pay to say
An ad intelligence API returns the ads a company is running, taken from the public ad transparency libraries platforms are required to publish. SocialCrawl covers the Facebook Ad Library in depth, with filters for country, status, media type and date range, plus Google ads by domain, on one API key and one shared schema.
Every one of these runs in production today. The endpoint chain under each card is the actual call sequence.
Search by company name to get the page, then request its ads. This two-step funnel is the most-used chain on the surface, and it works even when you only know the brand as a consumer would name it.
adlibrary/search/companies -> company/adsQuery the ad copy rather than the advertiser to find everyone bidding on a claim or a category term. This is how you discover the competitors you did not know were competitors.
adlibrary/search/ads (by keyword)Open a single ad for its full record, then request the transcript of its video. Turns a wall of creative into text you can classify, diff and search.
adlibrary/ad -> ad/transcriptPoint at a competitor's domain and get the ads it runs on Google. Pairs with the social lane to give you both halves of a media plan instead of one.
google/company/ads (by domain)The LinkedIn ad search runs on the same key. LinkedIn is the least stable upstream we operate, so treat it as an extra lane on a B2B report rather than the backbone of one.
linkedin/ads/searchRe-run the same advertiser query with a date range and record what changed. Creative that appears, disappears or shifts country is the signal, and it only exists if you kept yesterday's copy.
adlibrary/company/ads (date-ranged)Four calls take you from a company name to the creative and the words inside it. Start broad, then open only the ads that matter.
Reads are priced in credits. Empty result sets are not billed on list endpoints, and failed upstream calls are refunded automatically.
GET /v1/facebook/adlibrary/search/companiesSearch by company name to get the advertising pages behind it. Necessary because the page running the ads is often not named the way the brand is.
per callGET /v1/facebook/adlibrary/company/adsRequest the page's ads with filters for country, status, media type and date range. Cursor through for the full set.
per pageGET /v1/facebook/adlibrary/adFetch a single ad for its complete record, including the creative details you cannot see in a list response.
per callGET /v1/google/company/adsRun the same competitor domain against Google ads so the report covers paid search alongside paid social.
per callEach of these is a lane in the same competitive report. Follow a link to see that platform's full endpoint surface.
Most teams either buy a per-channel ad-spy seat or scrape the ad library themselves. Here is what changes.
| Capability | SocialCrawl | Typical setup |
|---|---|---|
| Access | Ad library data through a plain HTTP call, authenticated with one key. | A logged-in scraper that breaks when the page markup changes. |
| Filtering | Country, status, media type and date range are first-class query parameters. | Client-side filtering after pulling everything. |
| Creative depth | Single ad detail and video ad transcripts, so the creative becomes searchable text. | Thumbnails and a link, with the message left for a human to watch. |
| Coverage | Paid social and paid search on one key, alongside 46 platforms of organic context. | One channel per subscription, each with its own login. |
| Pricing model | Never-expire credit packs, priced per call, with refunds on upstream failures. | Monthly seats billed whether you ran a report or not. |
| Building a product on top | Permitted for transformed, customer-facing views. The terms say so in writing. | Usually restricted or negotiated case by case. |
An ad intelligence API returns the ads a company is currently or previously running, as structured records rather than screenshots. The source is the public ad transparency libraries platforms publish, so the data is the platform's own record of what was served.
The Facebook Ad Library in depth, covering company search, a company's ads, keyword search across ad copy, single ad detail and video ad transcripts. Google advertiser and domain ads sit on the same key. LinkedIn ad search is available and runs on our least stable upstream.
Yes. Country, status, media type, sort order and start and end dates are query parameters on the Facebook ads endpoints, and they are the filters production traffic actually uses. Results are cursor paginated so you can walk a full advertiser history.
No. You authenticate once with a SocialCrawl API key. There is no per-platform app to register, no review process and no separate credentials to rotate for the ad library.
Yes. Ad detail returns the full record for a single ad, and the ad transcript endpoint returns the spoken text of video creative where the library carries it. That turns a competitor's creative library into text you can classify and diff over time.
No. Public ad libraries publish the creative, the advertiser and the run dates, not the budget or the results. Anyone quoting exact competitor spend from library data is estimating. What you get here is what was actually run, which is the part that is verifiable.
Reads are priced in credits from never-expire packs, so you pay per call rather than per seat or per channel. A weekly competitor sweep is a handful of calls per brand. Signup includes 100 free credits and no card is required.
Ask AI about SocialCrawl
100 free credits on signup. No card, no sales call, no per-channel subscription.
curl -G "https://socialcrawl.dev/v1/facebook/adlibrary/search/companies" \
-H "x-api-key: $SOCIALCRAWL_API_KEY" \
--data-urlencode "query=Grow Therapy"curl -G "https://socialcrawl.dev/v1/facebook/adlibrary/company/ads" \
-H "x-api-key: $SOCIALCRAWL_API_KEY" \
--data-urlencode "pageId=123456789" \
--data-urlencode "country=GB" \
--data-urlencode "status=active"Both calls run against live data. Swap the key and the company and they work as written.