Skip to content

Dev-020/Orion

Repository files navigation

Orion: The Stateful, Multi-Modal AI Platform

Orion is a sophisticated, self-evolving AI designed for deep, continuous interaction. It combines advanced analytical capabilities with a distinct personality, functioning as both a versatile assistant and a collaborator in its own development.

Architecture

Orion operates on a modern Client-Server architecture orchestrated by a central TUI Launcher.

  • OrionCore (Backend): The brain of the operation. Built with Python, it handles state management, memory (SQLite + VectorDB), and LLM inference (API or Local).
  • Orion Server: A FastAPI-based middleware that exposes OrionCore functionalities via a REST API and WebSockets, enabling multiple frontends to connect simultaneously.
  • Launcher: A Text-Based User Interface (TUI) that orchestrates the entire system, managing processes for the Server, Discord Bot, Web Frontend, and GUI.

Key Components

  • Backend (backends/): Contains the core logic, memory systems, and API server.
  • Web Frontend (frontends/web/): A modern, responsive web interface built with React, Vite, and Tailwind CSS. Features include real-time chat, artifacts rendering, and user profile management.
  • Discord Bot (frontends/bot.py): A context-aware bot that brings Orion's personality to Discord, supporting multi-modal inputs and file handling.
  • Live Module (live/): An experimental module for real-time audio/visual interaction.

Key Features

  • Smart Launcher: A unified command center to start, stop, and monitor all services with live log viewing.
  • Authentication & Profiles: Secure user management with persistent profiles, avatars, and history privacy.
  • Multi-Modal Intelligence: Capable of processing text, images, and other file types seamlessly across all interfaces.
  • Self-Evolution: Equipped with tools to inspect and modify its own codebase, Orion can propose and implement features under operator supervision.
  • Hybrid Memory: Utilizes both structured SQL databases for factual consistency and Vector Databases for semantic context.

Getting Started

  1. Install Dependencies:

    pip install -r requirements.txt
    cd frontends/web && npm install
  2. Run the Launcher:

    python launcher.py

    Use the TUI to start the Server, Web Interface, and Discord Bot.

  3. Access the interfaces:

About

AI that will help me

Resources

Stars

Watchers

Forks

Contributors