Hannes Moser

On latest Chrome, the `<video>` element is referencing a MediaSource via a blob URL (e.g. blob:https://instagram.com/<ulid>). This is done by a proprietary video player (oz-player), which streams the video in chunks. For some browser versions (or IP's), a dash_init.mp4 URL is used.

The lengths Instagram (Meta) engineers go to in order to prevent scraping a simple web page are next level. I’m interested in creating reel and story transcripts and need to somehow get my hands on the video URL to extract the audio track. It’s not impossible, but boy, is it finicky.

This places feels nice.