We recently automated a task that used to take some manual hours and was pretty boring itself - documentation and user how-to instructions creation, since before the ability to use AI technology, this was pretty time-consuming thing to do. Now it takes around 15 minutes.

The workflow is simple: record your screen of a task, workflow or feature walkthrough (even without audio) -> get a step-by-step guide with titles, descriptions, and contextual screenshots out of it automatically, auto-translate, and edit the output if needed as in a regular doc before final export. How it works? by analyzing unique video frames via LLM and then combining that information into final docs.

This automates the process and makes documentation much less painful, because you can start with a solid base and refine if needed instead of a blank page. We built our own tool based on this workflow, and maybe will be useful for others, too.

It is free to try, and we are happy to hear any feedback! Thank you.