How to auto-generate developer documentation by analyzing streaming video: extracting text and speech with OCR and ASR, then structuring steps into readable markdown using LLMs.
Overview
At Augmend we’re developing ways to capture knowledge when devs work. We recently began experimenting with a new feature where we are able to auto-generate documentation just from “watching” your workflow. This allows for things like creating markdown for a how-to wiki without actually having to write anything or capturing the steps you took to solve a problem without taking notes yourself.
