Facebook Video & Reels Transcript Scraper

Get the transcript of any public Facebook Reel or video — plus caption, publish date, views, reactions, comments, duration, spoken language, a noSpeech flag and an optional AI summary & hook. Paste a Facebook page URL to transcribe its latest Reels and videos, or paste individual reel/video links. No Facebook login, no cookies, no API key needed.

Transcription runs on a built-in Whisper speech-to-text model inside the actor, so there is no extra AI bill. Want maximum accuracy? Plug in your own Groq or OpenAI key.

What can it do?

• 🎬 Facebook Reels to text – full transcript + optional timestamped segments (subtitles / SRT-ready)
• 📄 Bulk transcribe a whole Facebook page – latest N videos & Reels from any public page
• 🔗 Direct links – facebook.com/reel/…, /videos/…, /watch/?v=…, fb.watch and share links
• 📊 Engagement metadata – views/plays, reactions, comments & shares when Facebook shows them publicly
• 🌍 Language detection for 90+ languages
• 🔇 noSpeech flag – instantly filter music-only / silent clips
• 🤖 AI summary, hook & topics – with your own OpenAI / Anthropic / Gemini / OpenRouter / Groq key, or a free built-in heuristic
• 💸 Pay only per video – no subscription

Use cases

Content research & swipe files • competitor Reels analysis • repurposing videos into blogs/threads • hook libraries for UGC & ads • SEO & accessibility captions • social listening • training data for NLP.

Input

Field	Description
startUrls	Facebook page URLs and/or reel/video URLs
maxItems	Max videos in total (default 20)
maxItemsPerPage	Latest N videos per page
transcribe	Turn off for metadata only (cheaper)
transcriptionProvider	local (built-in Whisper, free), groq, openai
whisperModel	tiny (default, fast), base, small
maxTranscribeSeconds	Cut very long videos (default 600 s)
includeSegments	Add timestamped segments
llmProvider / llmApiKey / llmModel	Optional AI summary with your own key
directFirst	Connect without a proxy first, and use the proxy only after a block (default true)
proxyConfiguration	Fallback proxy, Residential by default

{ "startUrls": ["https://www.facebook.com/NASA", "https://www.facebook.com/reel/1803909017420449/"], "maxItems": 10 }

Output (one row per video)

{
  "url": "https://www.facebook.com/NASA/videos/artemis-acceleration/1803909017420449/",
  "videoId": "1803909017420449", "isReel": true,
  "pageName": "NASA - National Aeronautics and Space Administration", "pageUrl": "https://www.facebook.com/NASA",
  "caption": "We are not slowing down. 🚀 Our Artemis III mission will…",
  "publishedAt": "2026-10-23T23:37:19.000Z",
  "views": 584949, "reactions": 6600, "comments": 198, "shares": 507, "durationSeconds": 59.9,
  "transcript": "The standard that we should continuously aspire to…",
  "language": "en", "noSpeech": false,
  "hook": "The standard that we should continuously aspire to succeed…",
  "summary": "…", "topics": ["nasa", "artemis"], "aiMethod": "heuristic"
}

Export to JSON, CSV, Excel, or use the API, webhooks, Make, Zapier and n8n.

Pricing

Pay per event: video-transcribed per video with transcript, a cheaper video-scraped per video when transcription is off. See the price on the Pricing tab.

FAQ

Do I need a Facebook account? No. Only public content visible without login is scraped. Why are views/comments sometimes empty? Facebook doesn't show every counter to logged-out visitors on every video; fields are null rather than guessed. How accurate is the transcript? Built-in Whisper tiny is good for clear speech; choose base/small or a Groq/OpenAI key for noisy audio, accents or non-English. Is it legal? It collects only publicly available data. Respect personal-data laws (GDPR etc.) and Facebook's terms for your use case.

Keywords: facebook reels transcript, facebook video transcript, facebook video to text, fb reels scraper, facebook reels downloader transcript, speech to text, whisper.

Cost & proxy defaults

The default memory is 1 GB. That is enough for the tiny Whisper model, because the browser closes before transcription starts. Use 2 GB or more for base or small. Page listings load directly first, and the Actor switches to the fallback proxy (Residential by default) only when Facebook blocks the direct connection. Video files always download straight from Facebook's CDN.