Facebook Video Transcript Extractor

Facebook Video Transcript Extractor — Full Transcript and Video ID

Facebook Video Transcript Extractor turns any public Facebook video or Reel URL into clean, readable text. Paste one or more links and get back the full spoken transcript, the numeric video ID, and a collection timestamp for each video — structured JSON, ready to filter, export, or feed straight into an LLM pipeline. No login, no manual transcription, no video downloads. Run it from the Actor's Apify page and the first transcript lands in your dataset before the run finishes.

💡 What is Facebook Video Transcript Extractor?

Facebook Video Transcript Extractor is an Apify Actor that fetches one or more public Facebook video or Reel pages and returns the video's spoken-content transcript as plain text, along with its video ID and a scrape timestamp. It works directly against public URLs — no Facebook account, login, or cookies are used or required. It's built for content teams repurposing video into text, researchers and journalists working from video statements, and developers building searchable or LLM-ready video datasets.

📖 What Facebook video data is publicly available to scrape?

Anyone who can open a Facebook video's page in a browser without logging in can see its transcript (when the video has one), its URL, and its numeric ID — that's the same data this Actor reads.

Data category	Publicly available	Restricted
Video URL and numeric video ID	✅ Yes, for any public video/Reel link	—
Spoken transcript / captions	✅ Yes, when the video carries one	Not returned for private, friends-only, or region-restricted videos
Page/poster name, view counts, reactions, comments	✅ Visible on the page	Not extracted by this Actor — it returns transcript data only
Full video file / thumbnail	✅ Visible on the page	Not extracted by this Actor
Private or friends-only videos	❌ No	Requires a logged-in, authorized account

Facebook Video Transcript Extractor only returns publicly visible data — what any visitor sees. Nothing behind a login wall.

🧩 What data can I extract with Facebook Video Transcript Extractor?

Every run returns one row per video with five fields: the source URL, the video's ID, its transcript, an error message when extraction fails, and a scrape timestamp.

Identifying fields

Field Name	Description
url	The Facebook video URL you submitted.
videoId	The numeric Facebook video ID, extracted from /videos/<id>, /reel/<id>, or /watch/?v=<id> style links.
scrapedAt	ISO 8601 UTC timestamp of when the row was collected.

Content and status fields

Field Name	Description
transcript	The full spoken transcript text, cleaned of line breaks and extra whitespace. null if none could be found.
error	A short message explaining why no transcript was returned (e.g. no transcript exists, or the request failed). null on success.

🤖 Add-on: Need additional Facebook data?

If you also need the events a Page or Group is running, pair this Actor with Facebook Event Search Scraper. For review and recommendation posts left on Facebook Pages, use Facebook Reviews Scraper. Both live in the same Scraper Engine catalogue and pair naturally with video transcript data for a fuller picture of a Page's public content.

🔌 How does Facebook Video Transcript Extractor differ from the official Facebook API?

Meta's Graph API can return video captions, but only for videos your access token can see — arbitrary public videos posted by other people or Pages are out of reach without their authorization. As of 2026-08-15, Meta's own Graph API reference states you need "a valid Page Access Token or User Access Token" for any video, even one with public privacy settings, and the video-caption node returns metadata (locale, is_auto_generated, uri) pointing to a caption file rather than parsed transcript text (developers.facebook.com/docs/graph-api/reference/video/, developers.facebook.com/docs/graph-api/reference/video-caption/).

Feature	Facebook Graph API	Facebook Video Transcript Extractor
Access requirement	Page or User access token tied to the video, per Meta's own docs	Just the video's public URL
Video scope	Videos your token can access (owned, or ones you're tagged in)	Any public Facebook video or Reel URL, regardless of who posted it
Returned content	Caption file metadata and a uri link, not extracted text	Cleaned, plain-text transcript ready to read or index
Setup	Register a Meta developer app and generate an access token	Paste URLs into the input field and run
Output consistency	Follows the raw Graph API node/field structure	Consistent url / videoId / transcript / error / scrapedAt row every time

Use the Graph API when you own the Page or video and only need its own caption file. Use Facebook Video Transcript Extractor when you need readable transcript text for public videos you don't own, without registering an app or requesting a token.

🪄 How to use Facebook Video Transcript Extractor

Facebook Video Transcript Extractor runs entirely through the Apify platform — there's no separate signup or API key beyond your Apify account.

1. Open the Actor's page on Apify and start a new run.
2. Paste one or more public Facebook video or Reel URLs into the Video URLs (urls) field. The input schema doesn't mark this field as required, but the run stops with an error if it's left empty.
3. Optionally fill in Proxy Settings (proxyConfiguration) if you want automatic proxy fallback for unreliable requests.
4. Click Start and watch rows land in the dataset as each video finishes.
5. Download results as JSON, CSV, or another format Apify datasets support, from the Output tab.

How to scale to bulk transcript extraction

urls accepts an array, so a single run processes every video you list — there's no separate "bulk mode" to enable. The Actor loops through the list one video at a time, pushing a row per video as it finishes. If a paid run has a per-event charge limit configured, the Actor stops itself and logs "Charge limit reached, stopping the Actor" once that limit is hit, leaving any remaining URLs unprocessed for that run.

🎯 What can you do with Facebook video transcript data?

• 📝 Content teams repurposing video use the transcript field to turn a spoken statement into a blog post or caption draft without watching the clip.
• 🔬 Researchers and journalists use transcript alongside videoId and scrapedAt to build a timestamped, searchable archive of statements made across multiple videos.
• 🌐 SEO and accessibility teams use transcript to publish a text version of video content for indexing and screen readers.
• 🤖 AI engineers feed the transcript field of each row into a RAG pipeline or vector store, using url and videoId as stable document identifiers for citation and dedup.
• ⚙️ Developers monitor the error field to route failed rows into a retry queue instead of silently losing them.

🛡️ How does Facebook Video Transcript Extractor handle rate limits and blocking?

Requests go straight to Facebook's own pages and its api/graphql/ endpoint using standard HTTP requests with browser-like headers — there's no headless browser involved. Proxy handling follows a fixed fallback order, as implemented: a run starts with no proxy; if a request is refused and Proxy Settings is configured, it retries once through your configured proxy groups, then through a wider residential route if that also fails, retrying the residential route up to 3 times. Once a fallback tier succeeds, the Actor sticks with that tier for the rest of the run. If Proxy Settings is left empty, a failed request is not retried through a fallback tier. Each video is processed independently — a failure on one URL (page unreachable, no transcript found, request error) produces an error row for that video instead of stopping the run.

⬇️ Input

Parameter	Required	Type	Description	Example Value
urls	No	array (string list)	Paste one or more Facebook video URLs here. Each URL should point to a public Facebook video (e.g. facebook.com/.../videos/... or facebook.com/reel/...). Each link is processed and its complete spoken transcript extracted.	["https://www.facebook.com/100044280796927/videos/2301741263587145"]
proxyConfiguration	No	object (proxy editor)	Optional — configure proxy settings for this run to improve reliability. Leave empty to run without a proxy.	{"useApifyProxy": false}

Note: the schema doesn't flag urls as required, but a run started with an empty list fails immediately with 'urls' must be provided.

Example input

{
  "urls": [
    "https://www.facebook.com/100044280796927/videos/2301741263587145",
    "https://www.facebook.com/100044280796927/videos/4141502542759187"
  ],
  "proxyConfiguration": {
    "useApifyProxy": false
  }
}

⬆️ Output

Each processed video is written as one typed, normalized JSON row with the same five keys every time — a successful row carries a filled transcript and a null error; a failed row carries a null transcript and a short error message. Download results as JSON, CSV, Excel, or any other format Apify dataset exports support.

Successful rows are charged under the row_result pay-per-event; rows where no transcript could be extracted are pushed to the dataset directly and are not charged. To see only the charged, successful rows, filter your export where transcript is not null (equivalently, error is null).

Example output

[
  {
    "url": "https://www.facebook.com/100044280796927/videos/2301741263587145",
    "videoId": "2301741263587145",
    "transcript": "We're also working with the military on it because they want to make sure the technology holds up in the field before it ever reaches consumers.",
    "error": null,
    "scrapedAt": "2026-08-15T09:32:07.481203+00:00"
  },
  {
    "url": "https://www.facebook.com/100044280796927/videos/9999999999999",
    "videoId": "9999999999999",
    "transcript": null,
    "error": "No transcript found for this video",
    "scrapedAt": "2026-08-15T09:32:11.203981+00:00"
  }
]

⚙️ How does it work?

For each URL, the Actor requests the video's public Facebook page directly over HTTP. It reads the page's own embedded data — or, if needed, fetches a linked script file — to find the internal query identifier Facebook's front end uses for that video. It then calls Facebook's own GraphQL endpoint (api/graphql/) with that identifier and searches the response for the transcript text, stripping line breaks and extra whitespace. If the identifier can't be resolved dynamically for a given request, a last-resort fallback identifier is used so a transient lookup failure doesn't necessarily fail the video. No browser rendering is used. Only data visible on the public page is returned, and the output always keeps the same five-key shape regardless of how Facebook's internal page structure changes.

🔗 Integrations

Facebook Video Transcript Extractor runs on the Apify platform, so it works with anything that can call an Apify Actor or read an Apify dataset.

Calling it programmatically

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_API_TOKEN>")

run = client.actor("<your-username>/facebook-video-transcript-extractor").call(
    run_input={
        "urls": ["https://www.facebook.com/100044280796927/videos/2301741263587145"],
        "proxyConfiguration": {"useApifyProxy": False},
    }
)

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["videoId"], item["transcript"])

Works in Go, Ruby, Node.js, cURL — any language that can make an HTTP request.

No-code tools (n8n, Make, LangChain)

In n8n, use the HTTP Request node to call the Actor's run endpoint with your API token, then a second call against the dataset's run-sync-get-dataset-items endpoint to pull results. In Make, the HTTP module can drive the same two calls. In LangChain or LlamaIndex, use Apify's dataset loader integration to run the Actor and load transcripts directly into a document pipeline for retrieval or summarization.

⚖️ Is it legal to scrape Facebook video transcripts?

Scraping publicly available data — content any visitor can see without logging in — is generally lawful; the output of this Actor is limited to a video's URL, ID, transcript, and a scrape timestamp, not personal profile or account data. This falls under business/content-data scraping rather than a personal-data regime: Facebook's own Terms of Service govern what you may do with content collected from the platform, and database-rights considerations apply to any bulk collection you republish or redistribute. If a transcript captures a named individual's spoken statements and you intend to store or process it as personal data, consult legal counsel before doing so, since that use may bring GDPR or CCPA into scope regardless of the source being public.

❓ Frequently asked questions

What Facebook video fields does Facebook Video Transcript Extractor return?

Every row returns url, videoId, transcript, error, and scrapedAt — see What data can I extract above for the full description of each.

Does Facebook Video Transcript Extractor require a Facebook account or login?

No. The Actor sends unauthenticated HTTP requests with an empty cookie jar — it never logs in and never asks for Facebook credentials.

How many Facebook videos can I extract in one run?

As many as you list in urls — the Actor processes every URL in the array in one run, pushing one row per video, until the list is exhausted or a configured pay-per-event charge limit is reached.

What happens if a video is private or has no transcript?

The row is still returned rather than dropped: transcript is null and error explains why — for example "No transcript found for this video" if the video has no spoken-content transcript, or a request-failure message if the page couldn't be reached. Private and friends-only videos can't be read at all, since the Actor never authenticates.

Can I scrape multiple Facebook videos at once?

Yes. urls is an array — add as many video or Reel links as you need and each is processed and pushed to the dataset as it completes.

Does Facebook Video Transcript Extractor work with Reel URLs?

Yes. It recognizes facebook.com/.../videos/<id>, facebook.com/reel/<id>, and facebook.com/watch/?v=<id> link formats and extracts the video ID from any of them.

Does Facebook Video Transcript Extractor work with Claude, ChatGPT, and other AI agent tools?

It isn't exposed through an MCP server, but it's callable as a standard Apify Actor from any agent framework that can make an HTTP request or use the apify_client SDK — including custom tools built for Claude or ChatGPT agent frameworks.

Does Facebook Video Transcript Extractor return data in a format LLMs can use directly?

Yes. Every row is typed, normalized JSON with the same field names across runs — no HTML, no selectors to write. Pass transcript straight to an LLM, index it into a vector store, or feed it to an agent tool.

What happens when Facebook changes its layout or anti-bot system?

The Actor resolves its internal query identifier dynamically from each page rather than relying on a single hardcoded value, with a fallback identifier if that lookup fails for a given request — this is designed to absorb minor structural changes without failing every video. No specific update turnaround is promised.

Can I use Facebook Video Transcript Extractor without managing proxies or browser infrastructure?

Yes. It runs without a proxy by default and needs no browser setup on your side; enabling Proxy Settings adds automatic fallback through Apify Proxy if you want extra reliability on longer runs.

Which fields work best for AI training data and RAG indexing?

transcript is the high-information text field to index or summarize. url and videoId are stable, consistently structured identifiers to use as document keys or for deduplication across runs.

🔗 Related scrapers

Scraper Name	What it extracts
Facebook Event Search Scraper	Public Facebook events matching a keyword — title, schedule, venue, and attendance line.
Facebook Reviews Scraper	Review and recommendation posts from public Facebook Pages.
YouTube Video Subtitles Scraper	Subtitles/transcripts from YouTube videos in bulk.
YouTube Video Summarizer	YouTube video metadata, chapters, and timestamped transcripts, with optional AI summaries.

💬 Your feedback

Found a bug, or a video where transcript extraction fails unexpectedly? Let us know through the Issues tab on the Actor's Apify Console page so it can be looked into.