🛠️ Self-Hosted AI Stack via Docker Compose

The direction here is clear: packaging the mainstream local AI toolchain into a single Docker Compose deployment. It integrates Ollama, LiteLLM, and AnythingLLM, alongside voice models like Whisper and Kokoro, and even includes Docling and an MCP Gateway. In short, it tries to save you the headache of configuring these components one by one. The interesting part is its focus on being local-first and private by default. It supports NVIDIA CUDA acceleration and is compatible with both amd64 and arm64 architectures. For developers looking to quickly spin up a complete AI infrastructure on their own servers or NAS without wrestling with environment dependencies, this is a highly practical, out-of-the-box solution.