TikTok video transcript API for AI assistants

The TikTok video transcript API takes the link to a TikTok video and returns what is said in it as text. When the provider has them, each line comes with start and end times in seconds.

What it does

You pass the URL of a public TikTok video, and you can ask for a language. The provider uses the captions that TikTok already has when they exist, and runs speech recognition on the audio when they do not. The reply has the transcript and, for most providers, a list of timed segments.

The API reads public videos only. It does not download the video file for you, and it does not post or comment on TikTok.

How a Joshu assistant uses it

When you ask "what are people saying about our launch on TikTok", Joshu reads the transcripts of the videos that mention it and sorts what people said into praise, complaints, and questions.

When a creator pitches you a partnership, Joshu reads their recent videos and tells you what they talk about and how often they mention sponsors.

When you want to reuse one of your own videos, Joshu turns the transcript into a draft blog post or a newsletter section.

Choosing a provider

The main difference between providers is the timing data. Pick a provider that returns segments when you need to quote a moment or cut a clip, and the plain text endpoint when a summary is enough.

curl https://hello.joshu.me/api/router/call \
  -H "Authorization: Bearer $JOSHU_ROUTER_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "endpoint": "anyapi.tiktok.video_transcript_full",
    "params": {
      "url": "https://www.tiktok.com/@username/video/1234567890123456789"
    }
  }'