About

MiniMe MiniMe is a personal project developed by a group of enthusiasts. Our goal is to create personal agents trained on individual data without sharing that data with cloud systems or any other systems except those controlled by the owner. This ensures that both training and inference occur on user-controlled systems, utilizing local GPUs or specialized systems such as Nvidia Jetson or DGX platforms. Our goal is to go beyond what currently GPT systems are offering today and be user specific. Currently, we aim to release an open-source Alpha version by the third quarter of 2025 at the earliest.

Mini Me

MiniMe aims to be your digital twin, an agent that can serve as your personal assistant, impersonating you in day-to-day activities and improving over time by learning from you. Trained on personal data, including voice and written text, and augmented with industry-leading generative AI, MiniMe is designed to be uniquely you. The data used for training includes your social media activity, emails, and other written, audio, or video sources. Since it utilizes personal data, no cloud services are involved in the training process to ensure your privacy and security. We are developing secure connectors to extract your data for training purposes. This will generate the Base Model, which will be further refined using Reinforcement Learning based on your inputs. If you want to know more, please write to us at admin@minime.name