Helix is an open-source multi-model local LLM chat platform that enables collaborative AI discussions across different model runtimes (Ollama, LM Studio, llama.cpp).
- Multi-runtime support: Connect to Ollama, LM Studio, or llama.cpp servers via unified adapters
- 5 Group Chat Modes: Parallel responses, serial dialogue, role-based collaboration, and autonomous discussion
- Deep Monitoring: Real-time display of memory, GPU, parameters, quantization, and KV cache
- Parameter Tuning: Global, conversation-level, and model-level configuration
- Plugin System: Event-driven architecture for extensibility
- Cross-platform: Works on Linux, Windows, macOS
- Backend: Python (FastAPI + Socket.IO)
- Frontend: React + Vite + TypeScript
- Design: Anthropic Design System
- Storage: Local files (JSON/SQLite)
# Clone
git clone https://github.com/TbusOS/Helix.git
cd Helix
# Backend
cd backend
pip install -r requirements.txt
uvicorn helix.main:app --reload --port 8000
# Frontend (new terminal)
cd frontend
npm install
npm run devContributions are welcome! Please read CONTRIBUTING.md for guidelines.
Apache License 2.0
Helix 是一个开源的多模型本地 LLM 聊天平台,支持通过不同的模型运行时(Ollama、LM Studio、llama.cpp)进行协作式 AI 讨论。
- 多运行时支持:通过统一适配器连接 Ollama、LM Studio 或 llama.cpp 服务器
- 5 种群聊模式:并行响应、串行对话、角色分工、自主讨论
- 深度监控:实时显示内存、GPU、参数、量化类型、KV Cache 等信息
- 参数调优:全局、会话级、模型级三级配置
- 插件系统:事件驱动的可扩展架构
- 跨平台:支持 Linux、Windows、macOS
- 后端:Python (FastAPI + Socket.IO)
- 前端:React + Vite + TypeScript
- 设计:Anthropic Design System
- 存储:本地文件(JSON/SQLite)
# 克隆
git clone https://github.com/TbusOS/Helix.git
cd Helix
# 后端
cd backend
pip install -r requirements.txt
uvicorn helix.main:app --reload --port 8000
# 前端(新终端)
cd frontend
npm install
npm run dev欢迎贡献!请阅读 CONTRIBUTING.md 了解贡献指南。
Apache License 2.0