Facebook Video & Reel Transcript Scraper

Paste any Facebook video, Reel or Watch link and get the words back. Full transcript, timestamped segments, optional word-level timings, ready-to-save SRT and VTT subtitle files, the HD video URL, the thumbnail and the post's metadata. Whisper runs inside the Actor, so there is no API key to obtain and no proxy of your own to configure.

Input you paste	What you get back
facebook.com/reel/<id>	one row: transcript, segments, SRT/VTT, video + thumbnail URL, metadata
facebook.com/watch/?v=<id>	same
facebook.com/<page>/videos/<id>	same
fb.watch/<code>	same

Billed per second of audio, not per rounded-up minute. A 20-second Reel is billed 20 seconds.

How it works

1. Read the post. Every link is checked as a Facebook URL first — anything else is named and skipped rather than failing halfway through. The post is then read for its title, author, publish date, view count, thumbnail and the list of available video renditions.
2. Pull the audio. Only the audio track is downloaded; the video itself is never stored. Each attempt tries a direct connection first and falls back to the Actor's own residential pool where Facebook blocks datacenter IPs, so no proxy quota of yours is spent.
3. Transcribe. faster-whisper runs locally on the audio, VAD-filtered, and returns the language, the segments and — when you ask for them — per-word timings.
4. Package. Segments become SRT and VTT files, saved to the run's key-value store and linked from the row.

Input

Field	Type	Default	Notes
videoUrls	array	— (required)	Facebook video, Reel or Watch URLs, one per line
language	string	auto-detect	ISO 639-1 code (en, es, pt, de, …) to force a language
model	string	base	tiny (fastest) · base (fast, accurate on clear speech) · small (most accurate)
translateToEnglish	boolean	false	Return an English translation instead of the original language
wordTimestamps	boolean	false	Add per-word start/end/confidence inside every segment
maxMinutesPerItem	integer	120	Per-video cap; longer videos are transcribed up to the cap
cookies	string (secret)	—	Netscape cookies.txt, only for private or age-gated posts
proxyConfiguration	object	none needed	The Actor brings its own pool; set this only to add a fallback

Example input

{
  "videoUrls": ["https://www.facebook.com/reel/636216875539019"],
  "model": "base",
  "wordTimestamps": true
}

Output schema

One row per video. Real row from the run above, trimmed:

{
  "url": "https://www.facebook.com/reel/636216875539019",
  "title": "The viral spider video explained! 🤯",
  "uploader": "Steven Bartlett",
  "publishedAt": "2025-07-17",
  "viewCount": 3263423,
  "durationSeconds": 60.8,
  "transcribedSeconds": 60.8,
  "billedSeconds": 61,
  "language": "en",
  "languageProbability": 0.9962,
  "wordCount": 259,
  "text": "Have you ever seen that video of the ant in the pen? …",
  "segments": [
    { "start": 0.0, "end": 4.46, "text": "Have you ever seen that video of the ant in the pen?",
      "words": [ { "start": 0.0, "end": 0.26, "word": "Have", "probability": 0.84 } ] }
  ],
  "srt": "1\n00:00:00,000 --> 00:00:04,460\n…",
  "vtt": "WEBVTT\n\n00:00:00.000 --> 00:00:04.460\n…",
  "srtFileUrl": "https://api.apify.com/v2/key-value-stores/…/001-the-viral-spider.srt",
  "vttFileUrl": "https://api.apify.com/v2/key-value-stores/…/001-the-viral-spider.vtt",
  "videoUrl": "https://video.fsjj1-1.fna.fbcdn.net/o1/v/t2/f2/m366/…",
  "videoHeight": 1920,
  "thumbnailUrl": "https://scontent.fsjj1-1.fna.fbcdn.net/v/t15.5256-10/…",
  "error": null
}

Key output fields

Field	Description
text	Full transcript as one string
segments	[{ start, end, text }]; each also carries words when wordTimestamps is on
srt / vtt	Subtitle strings, ready to save
srtFileUrl / vttFileUrl	Direct download links to the stored .srt / .vtt
videoUrl / videoHeight	Highest-resolution video link Facebook listed, and its pixel height
thumbnailUrl	Post thumbnail
title, uploader, publishedAt, viewCount	Post metadata
durationSeconds / transcribedSeconds	Full length vs. what was transcribed after the cap
billedSeconds	Seconds charged for this video
language / languageProbability	Detected language and confidence
error	null on success; failed videos keep the reason and are never charged

Pricing

Event	When	Rate
Actor start	Once per run, per GB of run memory	$0.005
Second of media transcribed	Per second of audio actually transcribed	$0.0008

$0.0008/second works out to $0.048 per minute. A 20-second Reel costs about $0.026 all-in at the default 2 GB. Failed and skipped videos are never charged, and maxMinutesPerItem plus the run's maximum-cost setting bound the spend.

What makes this richer than the competition

Capability	Typical Facebook transcript Actor	This Actor
Billing granularity	per video or per rounded-up minute	per second of audio
Word-level timings	sometimes, extra cost	included, one toggle
SRT + VTT files	text only, or SRT only	both, as strings and stored files
HD video URL	rarely	included, with pixel height
API key needed	usually yes	no — Whisper runs in the Actor
Proxy quota needed	usually yours	no — the Actor brings its own pool
Languages	varies	99+, auto-detected

Notes & limitations

• Private, friends-only and age-gated posts need cookies. Public Reels and page videos work without them. Export a cookies.txt from a browser where the post opens normally and paste it into the cookies field.
• videoUrl and thumbnailUrl are signed CDN links and expire within hours. Download them during or shortly after the run; they are not durable storage.
• Live videos have to finish first. A stream that is still live or being processed cannot be transcribed; the row says so and nothing is charged. Retry once the recording is published.
• Longer videos need more run memory. Whisper decodes at roughly 2-3× realtime on the 2 GB default; for anything over about 20 minutes, raise the run's memory so it also gets more CPU.
• The model is clamped to the memory it fits in. small needs 2 GB; below that the Actor drops to a model that fits and says so in the log rather than being killed mid-decode.

🤖 For AI Agents & LLM Apps

• Purpose: Facebook video/Reel → transcript, subtitles and post metadata. One dataset row per input URL.
• Minimal tested input: {"videoUrls": ["https://www.facebook.com/reel/636216875539019"]}
• Output fields (flat): url, inputUrl, platform, sourceType, title, uploader, publishedAt, viewCount, thumbnailUrl, videoUrl, videoHeight, durationSeconds, transcribedSeconds, billedSeconds, billedMinutes, language, languageProbability, task, model, engine, wordTimestamps, text, wordCount, segments, srt, vtt, srtFileUrl, vttFileUrl, error.
• Nested: segments[] = {start, end, text, words?}; words[] = {start, end, word, probability}.
• Billing: media-second-transcribed × billedSeconds, $0.0008 each, plus apify-actor-start at $0.005 per GB of run memory. Bound spend with maxMinutesPerItem and the run's maximum cost.
• Behaviours worth knowing: non-Facebook URLs are skipped with a logged reason rather than failing the run; a video that cannot be fetched produces a row with error set and text: null and is not charged; the run stops cleanly and explains itself if it would otherwise hit its timeout.
• For other platforms (YouTube, TikTok, Instagram, X, podcasts, direct files) use memo23/video-audio-transcriber.

⚠️ Disclaimer

This Actor collects only data that is publicly accessible without logging in. It does not bypass authentication, and it does not collect private or personal data beyond what the post itself displays publicly. You are responsible for how you use the output, including compliance with Facebook's terms, applicable copyright, and data-protection law such as GDPR and CCPA. Transcripts are derived works of the source video — check the rights before republishing. Use for research, accessibility, subtitling and content analysis.

SEO Keywords

facebook transcript, facebook video to text, facebook reel transcript, facebook video transcription, facebook subtitle generator, facebook srt, facebook vtt, facebook watch transcript, fb.watch transcript, facebook speech to text, facebook video captions, facebook reel to text, whisper facebook, facebook video downloader url, facebook video metadata, transcribe facebook video, facebook accessibility captions, facebook content analysis