Facebook Ads Transcript Scraper

Find Meta/Facebook ads and export Meta-exposed captions plus reusable ad media URLs. The actor also returns creative copy, advertiser details, CTA data, and transparency metadata for competitive research.

This is an ad discovery and caption-detection tool—not automatic speech-to-text. Use returned videoUrls with an external ASR service when Meta does not publish captions.

What does Facebook Ads Transcript Scraper do?

Facebook Ads Transcript Scraper searches public Meta Ads Library pages and saves one dataset row per ad creative.

It captures:

• 🎬 Video ad URLs and preview images
• 📝 Creative body copy and headlines
• 🏷️ Advertiser/page categories and like counts when exposed
• 🔗 Landing page URLs, descriptions, CTA text, and CTA types
• 🧩 Creative format and collation metadata
• 📊 Spend, currency, and impression bands when available for political/EU transparency ads
• 🧾 Transcript, subtitle, closed-caption, speech, or voiceover fields when present
• 🧪 Diagnostics explaining whether transcript-like data was exposed

The actor is intentionally focused on creative analysis instead of broad political-ad archiving.

Who is it for?

Performance marketers use it to monitor competitor hooks and offers.

Creative strategists use it to build swipe files from live Facebook and Instagram ads.

Agencies use it to compare messaging across client categories.

Growth teams use it to discover videos that should be transcribed or reviewed by an editor.

Researchers use it to export public ad metadata without manually opening every ad card.

Why use this actor?

Meta Ads Library is public, but it is difficult to export repeatably.

This actor handles browser loading, proxy sessions, result normalization, and dataset output.

It also separates explicit Meta transcript/caption evidence from regular ad copy, so you know which creatives need external ASR. Millisecond caption timestamps are normalized to seconds and duplicate segments are removed.

Important transcript caveat

The actor does not invent audio transcripts.

It extracts transcript/subtitle/closed-caption/speech/voiceover fields only when Meta exposes them in the public Ad Library response.

If Meta only exposes the video URL and ad copy, the row will contain transcriptAvailable=false.

You can still use the videoUrls field to pass videos into a separate transcription workflow.

How much does it cost to extract Facebook ad transcripts?

Pricing uses pay per event.

• Start event: small one-time run fee
• Item event: charged per saved ad row

The default prefill is intentionally small so first runs stay cheap.

The current price is shown in the Apify Console before each run; item charges apply only to saved ad rows.

Input options

You can provide one or more of these inputs:

• searchQueries — keyword searches such as brand names, products, or hooks
• adsLibraryUrls — public Meta Ads Library result or detail URLs
• adArchiveIds — known numeric ad archive IDs

Search filters

The actor supports common Meta Ads Library filters:

• country
• activeStatus
• mediaType
• maxAds
• requireTranscript
• includeRawDiagnostics — defaults to true; set it to false to omit the concise diagnostics object

Use mediaType=video when your goal is transcript discovery. Existing Tasks remain compatible because diagnostics stay enabled by default.

Example input

{
  "searchQueries": ["fitness app"],
  "country": "US",
  "activeStatus": "all",
  "mediaType": "video",
  "maxAds": 25,
  "requireTranscript": false
}

Example output fields

Each dataset row may include:

Field	Description
adArchiveId	Meta Ad Library ID
pageName	Advertiser or Facebook page name
creativeText	Main ad body copy
title	Headline or card title
pageCategories, pageLikeCount	Advertiser metadata when exposed
caption, linkDescription	Creative caption/display domain and destination description
callToActionType, displayFormat	Machine-readable CTA and creative format
collationId, collationCount	Related creative grouping when exposed
currency, spend, impressions	Political/EU transparency metadata when exposed
videoUrls, videoPreviewUrls	Public media and preview URLs found in the response
transcriptAvailable	Whether explicit transcript/caption evidence was exposed
transcriptText, transcriptSegments	Meta-exposed text and normalized timestamped segments
transcriptSourceFields	Exact Meta response paths that supplied transcript evidence
diagnostics	Optional notes about extraction and missing transcript fields

Output quality

Rows are deduplicated by ad archive ID.

Dates are normalized to ISO date strings.

Arrays are used for media URLs because carousel/video ads can expose multiple assets.

Null values mean Meta did not expose that field for the specific ad. Spend and impression fields are commonly absent from commercial ads and are most often available for regulated political or EU transparency records.

Run summary and partial failures

Every run writes SUMMARY to the default key-value store. It contains successfulTasks, failedTasks, savedAds, and a per-input status/error list. In a mixed-input run, valid inputs still produce dataset rows while failed inputs remain machine-readable in this summary. If no usable rows are produced, the run fails with an actionable error instead of silently succeeding.

For archive-ID inputs, the actor only accepts the requested ID from Meta's response; it never substitutes an unrelated creative.

How to use it

1. Open the actor on Apify.
2. Enter competitor keywords, Ads Library URLs, or ad archive IDs.
3. Keep maxAds small for the first run.
4. Run the actor.
5. Export the dataset as JSON, CSV, Excel, or via API.
6. Filter rows where transcriptAvailable=true for transcript-ready creatives.
7. Send rows with only videoUrls to your ASR pipeline if needed.

Tips for better results

Use specific brand or product keywords.

Use activeStatus=active for current market research.

Use activeStatus=all for historical creative analysis.

Use mediaType=video for transcript workflows.

Disable requireTranscript if you also want video URLs and creative copy.

Enable requireTranscript only when you need a transcript-only dataset.

Integrations

Use this actor with:

• Google Sheets for competitor swipe files
• Airtable for creative review boards
• Slack alerts for new competitor ads
• BI dashboards for messaging trends
• Whisper or another ASR actor for videos without exposed transcripts
• LLM tools for hook, offer, and CTA classification

API usage with Node.js

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('automation-lab/facebook-ads-transcript-scraper').call({
  searchQueries: ['fitness app'],
  country: 'US',
  mediaType: 'video',
  maxAds: 10
});
console.log(run.defaultDatasetId);

API usage with Python

from apify_client import ApifyClient

client = ApifyClient('YOUR_APIFY_TOKEN')
run = client.actor('automation-lab/facebook-ads-transcript-scraper').call(run_input={
    'searchQueries': ['fitness app'],
    'country': 'US',
    'mediaType': 'video',
    'maxAds': 10,
})
print(run['defaultDatasetId'])

API usage with cURL

curl -X POST 'https://api.apify.com/v2/acts/automation-lab~facebook-ads-transcript-scraper/runs?token=YOUR_APIFY_TOKEN' \
  -H 'Content-Type: application/json' \
  -d '{"searchQueries":["fitness app"],"country":"US","mediaType":"video","maxAds":10}'

MCP usage

Claude Code

$claude mcp add --transport http apify-facebook-ads-transcripts https://mcp.apify.com/?tools=automation-lab/facebook-ads-transcript-scraper

Claude Desktop

Add this remote server in your MCP configuration:

{
  "mcpServers": {
    "apify-facebook-ads-transcripts": {
      "url": "https://mcp.apify.com/?tools=automation-lab/facebook-ads-transcript-scraper"
    }
  }
}

Cursor

Add the same server URL under Settings → Tools & MCP → New MCP server.

VS Code

Add the same server URL to your workspace MCP server configuration and enable the server before prompting your coding agent.

Example prompts:

• "Find active video ads for meal prep apps and summarize the hooks."
• "Export Facebook ad video URLs for this competitor keyword."
• "Show only ads where transcriptAvailable is true."

AI and external services

The actor does not call an AI model and does not perform speech-to-text. It accesses the public Meta Ads Library through Apify's execution and residential-proxy infrastructure. Meta receives the requested public search/detail request; no Meta login or user-supplied third-party API key is used.

Data handling and privacy

Inputs, public ad rows, the SUMMARY record, and operational logs remain in the user's Apify run storage and follow that account's retention settings. Browser cookies and proxy sessions are ephemeral to the run. The actor does not download video files, create an external cache, or send data to an AI provider. Users can delete runs and associated storage through Apify Console/API controls.

Legality and ethical use

This actor extracts publicly available Meta Ads Library data.

Respect Meta terms, privacy rules, and local laws.

Do not use scraped data for harassment, discrimination, or unlawful profiling.

FAQ

Does it always return a transcript?

No. Meta does not expose transcript-like fields for every video ad. The actor returns transcriptAvailable=false when only video URLs and creative copy are available.

Can I use the video URL with another transcription tool?

Yes. Use videoUrls for downstream ASR workflows when Meta does not expose transcript text directly.

Troubleshooting: no results

If a query returns no rows, try a broader keyword, a different country, or activeStatus=all.

Facebook may also temporarily block or throttle a session.

Troubleshooting: no transcripts

Many video ads expose only the video URL and ad copy.

That is normal.

Use transcriptAvailable and diagnostics.extractionNotes to understand what happened.

Related scrapers

Related automation-lab actors:

• https://apify.com/automation-lab/facebook-ads-library
• https://apify.com/automation-lab/meta-ad-library-page-resolver
• https://apify.com/automation-lab/google-ads-transparency-center-scraper
• https://apify.com/automation-lab/pinterest-ads-library-scraper

Best practices

Start with 10 to 25 ads.

Validate output fields.

Increase limits only after confirming the query returns useful creatives.

Run separate jobs for separate competitors so datasets stay easy to compare.

Data retention

Apify stores run datasets according to your account settings.

Export or integrate important results into your own systems.

Support

If a public Ads Library URL fails repeatedly, include the run ID and input when reporting the issue.

Changelog

Initial version: keyword, Ads Library URL, and ad archive ID inputs with creative metadata, media URLs, transcript availability flags, and diagnostics.