Second Me - Mind Universe's open-source AI identity model

What is Second Me?

Second Me is an open-source AI identity model developed by Mindverse. It supports the creation of fully private and deeply personalized AI agents that represent a user's "true self." Second Me offers two interaction modes: Chat Mode and Bridge Mode. Chat Mode supports personalized dialogue, while Bridge Mode acts as a bridge connecting the user to the world, enabling personalized feedback and enhancement of information. Second Me can run locally, ensuring absolute privacy of user data. Second Me helps users express themselves flexibly in different contexts, allowing them to regain control of their identity and data in the AI era.

Second Me's main functions

• Personalized identity creationUsers upload their memories, experiences, and preferences to train an AI agent that can represent the user's true self.
• Multi-role adaptationAutomatically switch roles based on different scenarios (such as work, social, and study) to maintain the user's core identity.
• Chat ModeIt allows users to engage in direct dialogue and provides personalized responses based on their personal memories.
• Bridge Mode: Serving as a bridge between users and the outside world, enhancing the expression of needs and information feedback.
• Privacy protectionSecond Me operates entirely locally, with user data stored on the local device.
• Intelligent memory managementIt supports rapid pattern recognition, adapts to changes, and evolves together with users.

Second Me's technical principles

• Hierarchical Memory Model (HMM):
  • L0 (Short-Term Interchangeable Memory)It handles immediate contextual information, used for short-term interactions and quick responses.
  • L1 (Natural Language Memory Layer)It summarizes and stores important user information, such as personal profiles and preference tags.
  • L2 (AI native memory layer)It learns and organizes memory based on model parameters to perform complex reasoning and knowledge retrieval.
• Me-alignment architectureBased on reinforcement learning technology, it transforms users' scattered data into a deep, personalized understanding, ensuring that AI accurately grasps users' preferences and behavioral patterns.
• Decentralized networksEach Second Me is an independent AI entity that communicates and collaborates based on a peer-to-peer network, ensuring data privacy and security.
• Automated training pipelineThis includes steps such as data synthesis, filtering, supervised fine-tuning (SFT), and direct preference optimization (DPO) to ensure efficient model training and personalization.
• Multi-agent frameworkIt supports collaboration with other AI agents or expert models, providing users with more accurate services based on enhanced contextual information and optimized interaction processes.
• Chain reasoningBased on the CoT style during training and inference, step-by-step reasoning and detailed explanation are used to improve the model's logic and accuracy.

Second Me's project address

• Project official website:https://www.secondme.io/
• GitHub repository:https://github.com/mindverse/Second-Me
• arXiv technical paper:https://arxiv.org/pdf/2503.08102

Second Me application scenarios

• Personal AssistantIt helps manage schedules, emails, and tasks, saving time and increasing efficiency.
• Career DevelopmentAnalyze resumes and skills, provide job search advice, and simulate interview scenarios.
• Social interactionProvide social suggestions based on user style, and even reply to messages on their behalf.
• Learning tutoringProvides personalized learning resources, organizes notes, and answers questions.
• Health supportRecord health data, provide health advice, and remind users to take medication and exercise.