A desktop application that handles your emails securely and informs you if there are any important emails. THe idea is to cut the time it takes us for checking our emails periodically and more importantly none of the important emails are missed.
I worked on EmailAIAssistant, a desktop application which used local LLAMA models using OLLAMA and notifies you of any important incoming email.
Electron - Using Electron for building cross platform desktop app.
LLAMA with OLLAMA - Running LLAMA on OLLAMA on desktop to make sure that sensitive email data is never exposed to third party servers. Also LLAMA is used to create a long term memory for the user's preferences.
React - The UI is built on React with Chakra UI to deliver a modern clean interface.
Gmail Integration - The app pulls the emails form Gmail APIs to process them in realtime.
Long Term Memory - The system keeps a long term memory of the chats, and had ability to perform filtration based on this long term memory.
Chat Interface - Using chat interface simplifies the training of the LLM for the desired behaviour.
This application solves many challenges, starting with building cross platform desktop application, training LLAMA model on the fly and processing email in realtime.
EmailAIAssistant delivers a very secure way to process the emails in realtime and makes sure that filtration is highly customizable and realtime.