Video Transcript Scraper: TikTok, Instagram, YouTube & X
Pricing
from $1.60 / 1,000 transcript from subtitles
Get transcripts from TikTok, Instagram Reels, YouTube videos and Shorts, and X videos. Native subtitles first, Whisper speech-to-text when a video has none. Timestamped segments, language, author, caption and stats. Pay only for transcripts.
TikTok videos, Instagram Reels and video posts, YouTube videos and Shorts, and X (Twitter) posts with a video. Short links (vm.tiktok.com, youtu.be) work too. One transcript row per video. Provide video URLs, profiles below, or both.
Get transcripts for the latest videos of a creator: tiktok.com/@name, instagram.com/name, youtube.com/@channel (add /shorts for the Shorts tab). Instagram serves at most the 12 newest posts without a login.
How many of the newest videos to transcribe from each profile or channel.
When to transcribe the audio with Whisper. Subtitle transcripts cost $2 per 1,000 and speech-to-text transcripts $5 per 1,000 (first minute of audio included, then $0.0025 per started minute), at list price.
Fallback (default):the platform's own subtitles when a video has them (TikTok, YouTube), speech-to-text otherwise (always for Instagram and X).
Replace automatic captions: like Fallback, but videos whose captions are machine-made are transcribed with speech-to-text and billed at the speech-to-text price. In our test of 500 videos, automatic captions missed about 1 meaningful word in 120, and 5% or more of the words in 7% of videos; YouTube's also censor swearing. Captions written by the creator are kept. If speech-to-text fails, you get the captions at the subtitle price.
Always: speech-to-text for every video, at the speech-to-text price.
Never: subtitles only; videos without them are skipped for free.
Two-letter language code such as en, es or fr. Picks the subtitle track in that language when a video has several, and tells speech-to-text which language is spoken. Leave empty to detect it automatically.
Add the transcript split into segments with start and end times in seconds. Turn off to get the plain text only.
Videos longer than this are not sent to speech-to-text (they still get subtitles when they have them). Protects you from paying for a 3 hour livestream by mistake.
