Engines

Ollama, LM Studio, llama.cpp, vLLM и Triton: обзор движков инференса LLM

The easiest way to build with open models

AI that lives on your computer. Open-source, private, always local.

For multy-GPU servers

A terminal-based coding agent