Instagram post transcript API for AI assistants

The Instagram post transcript API takes the link to a reel or a video post and returns the words spoken in it as text, along with the caption and the length of the video. Your assistant can then read a two minute reel in the time it takes to read a paragraph.

What it does

You pass the URL of a public Instagram post. The provider fetches the video, runs speech recognition on the audio, and returns the transcript in the language of the video. Posts with no speech come back with an empty transcript and the caption only.

The API works on public posts. It cannot read posts from private accounts, and it does not post, like, or comment.

How a Joshu assistant uses it

When a client sends you a reel and asks what you think, Joshu reads the transcript and writes a three line summary in the reply draft, so you can answer without watching it on your phone.

For a research report on a topic, Joshu collects the reels that people share about it and quotes what was said, with a link back to each post.

Choosing a provider

Providers return the same transcript in different shapes. Some return one block of text, and others return segments with timestamps, which help when you want to quote the exact moment a point was made.

curl https://hello.joshu.me/api/router/call \
  -H "Authorization: Bearer $JOSHU_ROUTER_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "endpoint": "auto",
    "capability": "instagram.media.transcript",
    "params": {
      "url": "https://www.instagram.com/reel/abc123/"
    }
  }'