Instagram Transcript Extractor - Reel Speech to Text, SRT & VTT
Pricing
from $8.42 / 1,000 results
Instagram Transcript Extractor - Reel Speech to Text, SRT & VTT
Turn any public Instagram Reel into a clean transcript: full text plus timestamped segments, SRT and VTT subtitles, in the spoken language (auto-detected). Real speech-to-text with Whisper. You only pay for Reels that return a transcript. $9.90 per 1,000.
Pricing
from $8.42 / 1,000 results
Rating
0.0
(0)
Developer
Alex Vega
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
an hour ago
Last modified
Categories
Features
Share
Instagram Transcript Extractor
Turn any public Instagram Reel into a clean transcript: full text plus timestamped segments, with SRT and VTT subtitles, in the spoken language (auto-detected, Spanish included).
How it works
Most Reels have no caption track, so this Actor does real speech-to-text: it downloads the Reel's audio and transcribes it locally with OpenAI's Whisper model (your choice of small for accuracy or base for speed). No login is used and no Reel is ever charged unless it actually returns a transcript.
Input
- Reel URLs — public Reel or video-post URLs (one per line).
- Max Reels — overall cap for the run.
- Whisper model —
small(default, better with accents and Spanish) orbase(about 2x faster). - Spoken language — leave empty to auto-detect.
- Proxy — datacenter by default (verified working); residential available.
Output (one row per delivered transcript)
url, shortcode, author, caption, publishedAt, durationSeconds, language, transcript, segments ({startMs, durationMs, text}), srt, vtt, wordCount, plus like/comment counts at extraction time.
Pricing
Pay per result: $9.90 per 1,000 transcripts. Reels that are private, unavailable, or contain no speech (music-only) are classified in the run log and never charged.
Honest limits
- Public Reels only; private or age-gated media cannot be fetched without an account, and this Actor does not use accounts.
- Music-only Reels return no transcript (and are not charged).
- Speech-to-text is automatic: names and rare words can be misheard, more so with the
basemodel.
