Screenshot to GPT - Watchtower Introduction

This project uses an Dynamic Express Server to power reponses from a local instance. The local instance listens to active image changes under the /media folder.

The local instance then converts the image to text using python tesseract, sends converted text to OpenAI using the specified model and sends that response to the express server, using endpoint api/watchtower/

The watchtower actively listens to the latest reponse to the api/watchtower endpoint and places the text in html using Typescript.

Video