Instagram Transcript API

Instagram Transcript

Features

• Multiple source URLsSubmit one or more public Instagram post or Reel URLs and receive transcript records for media that exposes spoken content.
• Carousel-aware resultsPreserve separate transcript records when one carousel post produces more than one media transcript.
• Normalized text and provenanceReturn transcript text with language, plaintext or WebVTT format, source URL, media identifier, shortcode, and collection time.

Parameters

How to Use

Submit public Instagram media URLs as an asynchronous task, poll its task ID, then read every normalized transcript record with cursor pagination.

1. Prepare one or more public Instagram post or Reel URLs.
2. POST the urls array to /v1/instagram/transcript.
3. Poll /v1/tasks/{task_id} until the status is succeeded or failed.
4. Read /v1/tasks/{task_id}?limit=50, reconcile zero or multiple records per input URL, and follow results.next_cursor while results.has_more is true.

curl -X POST "https://api.socq.ai/v1/instagram/transcript" \
  -H "Authorization: Bearer $SOCQ_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"urls":["https://www.instagram.com/reel/DHsD6HGqJhp/"]}'

# Poll GET /v1/tasks/{task_id}
# Read data.results.items and follow next_cursor while has_more is true

Best Use Cases

• Spoken-content search: Index transcript text with language, source URLs, and shortcodes so selected public posts and Reels can be searched by spoken terms.
• Post and Reel research: Review transcript text across a curated URL set while retaining media identifiers and collection time for source checks.
• Caption and accessibility review: Use returned text and format metadata to prepare readable copies or inspect whether selected public media has usable spoken-content records.

Pro Tips

• Submit canonical public post or Reel URLs and remove exact duplicates before starting a task so each intended source is processed once.
• Reconcile results by record ID rather than URL alone because silent media can return no record and one carousel URL can return several.
• Branch downstream parsing on format; preserve WebVTT cues when timing structure matters and treat plaintext as unsegmented text.
• Use collected_at for collection lineage because created_at is not populated with the Instagram media publication time.
• When chunking transcript text, retain the record ID, source URL, shortcode, language, format, and collected_at with every derived segment.

Related APIs