momori777/Artemis

破限本地AI女友后宫，Agent+画图语音向量数据库+live2D+桌宠+酒馆角色卡导入+webchat，QQ+Telegram双通道，8G显存可跑 uncensored Fully offline AI girlfriends harem Openclaw/Claude code+Local LLM+GPT-SoVITS+ComfyUI image+Live2D+desktop pet+SilllyTavern Character card import+frontend | QQ & Telegram | Dynamic 8G VRAM scheduling+mem0 qdrant, can run offline

What it solves

This project provides a fully local, private, and uncensored AI companion system. It eliminates the need for cloud-based AI services, ensuring that all conversations, voice synthesis, image generation, and character animations remain entirely on your own machine without data leakage or censorship.

How it works

It integrates several open-source technologies into a unified local daemon. It uses llama.cpp for large language model inference, GPT-SoVITS for voice synthesis, and ComfyUI for image generation. These components are orchestrated by a central agent framework that manages character memories, relationship states, and desktop pet animations (via Live2D). The system supports hot-swappable characters, each with isolated memory and personality profiles, and can be accessed via a web interface, QQ, or Telegram bots.

Who it’s for

This project is for users who want a highly customizable, private AI companion and have the hardware (specifically NVIDIA GPUs) to run multiple local models simultaneously. It is designed for those who want to avoid subscription costs, cloud-based censorship, and privacy risks associated with commercial AI girlfriend platforms.

Highlights

  • 100% Local Execution: Zero API dependencies; all processing happens on your hardware.
  • Multi-Character Support: Hot-swappable characters with isolated long-term memories and unique personality traits.
  • Comprehensive Integration: Combines LLMs, TTS, image generation, and Live2D desktop pets into a single cohesive experience.
  • Advanced Relationship Engine: Features a layered behavior engine that tracks interest, trust, and conflict levels to influence character responses.
  • Extensible Architecture: Allows users to import character cards from other platforms and swap between different LLM and image generation models.

Related

  • Project
    ls1intum/Artemis

    Artemis is a self‑hosted, MIT‑licensed platform for university‑level interactive learning. It supports many exercise types, automatic grading, built‑in version control/CI, plagiarism detection and adaptive learning. Optional AI services (Iris, Athena, Hyperion, Atlas agents) provide LLM‑based tutoring, feedback suggestions and authoring assistance. Deployable via a single Docker‑Compose command, it’s used by multiple institutions and released bi‑weekly.

  • Project
    google/artemis

    An AI-powered Android automation framework that enables AI agents and test suites to control real phones using natural language, achieving SOTA results on the AndroidWorld benchmark.

  • Project
    fathah/hermes-desktop

    Hermes Desktop is an Electron‑based GUI that installs, configures, and lets you chat with the Hermes Agent LLM assistant. It supports local or remote back‑ends, dozens of LLM providers, a streaming chat UI with slash commands, profile & session management, tool integrations, scheduling, and many messaging‑gateway bots.

  • Project
    PunithVT/ai-avatar-system

    AvatarAI is an open‑source, production‑grade system that turns any portrait photo and a short voice sample into a real‑time, lip‑synced talking avatar. It streams Whisper‑transcribed speech to a selectable LLM (Claude, GPT‑4o, or local Ollama/vLLM), feeds the generated text to a multilingual TTS (Chatterbox/XTTS), and renders video frames with MuseTalk V1.5. The stack includes Next.js front‑end, FastAPI back‑end, PostgreSQL, Redis, Celery, Prometheus, and Docker‑Compose, with optional one‑click AWS GPU deployment. Features include zero‑shot voice cloning, barge‑in interruption, multi‑LLM support, JWT auth, conversation memory summarisation, and optional face‑restoration via GFPGAN.

  • Dispatch
    AI & Frontier Tech Roundup – Qwen 3.8, Agentic Coding, and Emerging Governance

    Qwen 3.8 dominates local AI benchmarks, agentic coding frameworks mature, and new governance pressures reshape model deployment.