Facebook Video Transcript API

Facebook Video Transcript

Features

• Known-URL transcript batchesAccept a list of public Facebook video post or Reel URLs and remove exact duplicate strings before transcript retrieval.
• Source-linked transcript recordsReturn each available transcript with its submitted source URL and a record ID that falls back to that URL.
• Explicit text formatIdentify returned text as plaintext or WebVTT and include the transcript collection time for downstream handling.
• Available-only resultsCreate a normalized record only when a submitted public URL yields non-empty transcript text, so result counts can be lower than input counts.

Parameters

How to Use

Submit known public Facebook video post or Reel URLs, then reconcile returned source URLs because only non-empty transcript text produces result items.

1. Prepare one or more known public Facebook video post or Reel URLs; this endpoint does not discover a Page video feed.
2. POST urls to /v1/facebook/video-transcript and store the returned task_id.
3. Poll /v1/tasks/{task_id} until the task succeeds or fails.
4. Read data.results.items, follow next_cursor while has_more is true, and use each url and format when storing text.

curl -X POST "https://api.socq.ai/v1/facebook/video-transcript" \
  -H "Authorization: Bearer $SOCQ_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"urls":["https://www.facebook.com/reel/1535656380759655"]}'

# Poll GET /v1/tasks/{task_id}
# Read GET /v1/tasks/{task_id}?limit=50

Best Use Cases

• Spoken-content search indexing: Index transcript text with its source URL, record ID, format, and collection time so selected public videos can be searched by spoken words.
• Selected video and Reel review: Use source-linked transcript text to review the spoken material in a curated set of known public video post and Reel URLs.
• Post metadata enrichment: Join transcript text to separately collected post details by source URL when a dataset needs both spoken content and public post context.

Pro Tips

• Submit known public video post or Reel URLs after checking that each target is reachable; malformed, private, removed, or unsupported media can produce no record.
• Canonicalize URL strings before batching when the same video can appear with tracking parameters or alternate URL forms, because the backend removes exact duplicates only.
• Compare returned url values with the submitted list after success so URLs without non-empty transcript text remain visible to your workflow.
• Read format before parsing text; a webvtt record can retain cue markup that should not be processed as unstructured plaintext.
• Continue with results.next_cursor while has_more is true when a URL batch spans more than one result page.

Related APIs