Meta Ad Library API speed test: vexqo and Apify's actor
On 20 November 2026 we ran four matched jobs on vexqo and on Apify's official Facebook Ads Library Scraper. vexqo answered sooner in all four, and both returned the same ads. The test is small. This page gives the numbers and what the test does not show.
Four jobs, timed on both services
The keyword jobs searched "running shoes" among active ads in the United States, ordered by impressions. The 30-ad search ran three times on each service, and the other jobs ran once.
| Job | Runs on each | vexqo | Apify actor | Difference |
|---|---|---|---|---|
| Keyword search, 30 ads | 3 | 3.0 s median, from 2.9 to 3.7 s | 61 s median, from 18 to 61 s | 20.3 times, by the medians |
| Keyword search, 100 ads | 1 | 14.1 s | 54 s | 3.8 times |
| One advertiser's page, 30 ads with page information | 1 | 8.3 s | 34 s | 4.1 times |
| One ad with its details | 1 | 4.4 s | 15 s | 3.4 times |
Both services returned the same ad IDs in every keyword search, and all six requests on each succeeded. vexqo was timed as one request to a warm service, and Apify as the run duration its console shows. The actor is Apify's Facebook Ads Library Scraper, build 0.0.410 with 1 GB of memory.
What this test does not show
Twelve requests in eight minutes are a first measurement. These two claims would need a wider test.
A speed guarantee
The test used one keyword, one country, one advertiser's page and one ad. Other searches and other hours can differ.
Reliability
Every request succeeded on both services. Six requests each cannot support a success rate or a slowest-case figure.
Questions
Something else? Write to [email protected].
Who needs an answer this fast?
Products where a person or an AI agent waits for the result: a competitor lookup inside a dashboard, or an assistant that searches ads in the middle of a conversation. For a batch that runs overnight, the difference matters less.
Why is the 30-ad difference so much larger than the others?
Apify's three runs took 61, 18 and 61 seconds, so its median is 61. An actor run includes its own start, and we did not separate that from the time spent reading ads. The range is in the table for that reason.
Will you run a wider test?
A wider test needs several keywords, countries and times of day, the same outside client for both services, and cold and warm starts kept apart. Its results will go on this page when it is run.
Is vexqo an official Meta API?
No. vexqo reads the public Ad Library, logged out, with no Meta account. It is web scraping, with no approval from Meta and no claim of one. Check that your intended use fits their terms.
What does a vexqo ad record hold?
The Meta Ad Library API page shows a real ad record. What Meta's official API returns covers the official route.
Time it yourself
The trial is free and has 500 credits, enough for about 5,000 ads. Run the same 100-ad search from your own machine and time it.
Get an API key