AlphaAvatar/AlphaAvatar

A real-time interactive Omni Avatar built on LiveKit, which allows you to seamlessly integrate with any open source Avatar components (real-time model, visual, voice, memory, search, etc.).

What it solves

AlphaAvatar provides a framework for building self-hostable, privacy-first personal AI assistants that act as "intelligent butlers." It moves beyond simple chatbots by enabling agents to be stateful, proactive, and multimodal, allowing them to remember user preferences, plan long-term tasks, and interact across various channels like voice, text, and video.

How it works

The system uses a plugin-based real-time agent architecture that processes multimodal observations (voice, camera, screen) through a core perception layer. It integrates several key modules:

  • Memory & Persona: Uses graph-aware multimodal memory and automatic persona extraction to maintain identity continuity and user understanding.
  • Interaction Router: Manages real-time speech routing, turn-taking, and interruptions.
  • Tool Integration: Connects to the external world via MCP (Model Context Protocol), RAG for document retrieval, and DeepResearch for web searching.
  • Virtual Character: Provides a visual representation of the assistant for real-time interaction.
  • Infrastructure: Connects to various LLM providers and storage backends (like LanceDB or Qdrant) for vector and identity data.

Who it’s for

It is designed for developers and users who want a highly personalized, self-hosted AI companion capable of managing personal life metrics, organizing knowledge, scheduling tasks, and executing autonomous multi-step workflows.

Highlights

  • Privacy-First: Fully self-hostable, giving users total control over their data and memory.
  • Multimodal Interaction: Supports real-time voice, text, and camera inputs with a dedicated virtual character interface.
  • Proactive Planning: Includes modules for task decomposition, reminders, and long-horizon goal execution.
  • Extensible Tooling: Integrates with MCP for external APIs and RAG for personal knowledge bases.
  • Diverse Access Channels: Available via a web demo, LiveKit playground, and messaging platforms like WhatsApp.

Related

  • Project
    KevinLiss/ApeAdmin

    ApeAdmin is an open‑source FastAPI + Vue 3 admin framework tailored for AI applications. It offers a WordPress‑style installation wizard, full RBAC, plug‑in architecture, and an MCP‑SSE gateway that turns back‑end functions into tools callable by LLM agents. Built‑in multi‑model chat, audit logs, and a plug‑in marketplace let you quickly build a secure management UI and expose business logic to AI agents without writing extra APIs.

  • Project
    RTGS2017/NagaAgent

    NagaAgent is a cross‑platform desktop AI assistant (Windows/macOS/Linux) that offers chat with any OpenAI‑compatible LLM, custom tool‑calling via JSON blocks, graph‑RAG memory stored in Neo4j, a Live2D animated avatar, voice synthesis/recognition with lip‑sync, a plug‑in MCP framework for weather, web search, game guides, etc., and community features like a forum and skill marketplace. It ships as an Electron/Vue front‑end plus Python FastAPI back‑end, is installable via `uv sync` (or pip), and is dual‑licensed (AGPL‑3.0 + proprietary).

  • Project
    dromara/wgai

    WGAI is a Spring‑Boot + Vue web platform that bundles vision (OpenCV, YOLO, OCR, face ID), audio (speech‑to‑text, ChatGPT‑style dialogue), digital‑human avatars, and AGV navigation into an offline‑first, industrial‑grade AI management system. It offers online data annotation, model training, real‑time video/audio analysis, and system monitoring, with a UI for configuration and a public demo instance. The code is AGPL‑3.0 and support is mainly via a paid Chinese community.

  • Dispatch
    ChatGPT Plugins Release

    OpenAI has introduced plugins for ChatGPT, allowing the model to access real-time information, perform computations, and integrate with third-party services to expand its utility and reduce hallucinations.

  • Project
    lc2panda/alphastream

    Alphastream is an AI‑driven financial analysis web app. It uses DeepSeek/any OpenAI‑compatible LLM with LangGraph to run 14 specialized agents (technical, fundamental, sentiment, risk, etc.) plus four investor‑persona agents. Real‑time market data comes from dual adapters (AKShare & BaoStock) covering Chinese, US, Hong‑Kong, crypto and macro sources. The Next.js + React front‑end shows a dark glass‑morphism UI with chat, dashboards and chart cards, while the Flask backend streams agent progress via SSE. All services can be launched with a single Docker‑Compose command.