Delphi allows users to speak to their computer and have it complete actions for them. It does this via function calling using either xLam or gpt-4o. The agent also has a preliminarly feature that allows it to write and execute code(so use this repo at your own risk). The goal is to be useful to humans because speech is much more vibrant and natural for humans than keyboards and mice. Especially for older people and those with disabilities, speaking with the computer is often easier than the current options. This project is not actively maintained anymore because better alternatives exist.
run pip install -r requirements.txt
Ongoing list of dependencies, but the current working product is in 'pipeline_OpenAI.py' which requires a few things. 1. You need to have ollama installed and 2. You need to put an OpenAI api key in a .env file. You should expect it to take some time to run, but once it is setup you can type a query and it should do something.
By Adit Kadepurkar, Matthew Xu