A developmental AI project to turn any text passage into notes (mainly so I can catch up on my AP World notes faster)
From Text2Notes's About page:
"The technology at the heart of this project is Python code which processes text input into a feedable form for OpenAI's text-davinci-003 model from the GPT-3 family via API. This crude website serves as an interface between the technology and its users, written with the Flask framework.
To use, copy text from any source, preferably from a published textbook, into the left-hand textarea and submit. The process of conversion is indicated by the "Submitted!" text, a time at which no further input should be applied to the site. Depending on the length of input, converted notes form should appear in the right-hand textarea in a timely manner. Issues can be reported directly on this project's Github page.
Due to OpenAI's current restrictions on its API, queries for the text-davinci-003 model are limited to 4096 tokens, or about 3000 words. In correspondance with the nature of this technology, this means that individual paragraphs must be limited to under 3000 words. In addition, use of the API costs real money, meaning that with each conversion my wallet loses weight. While the Beta release offers the tool free of charge and will remain this way for the forseeable future, I would appreciate it if spamming, overuse, and abuse are brought to a minimum."