ollama / ollama

Runs open-weight LLMs such as Gemma, Qwen, DeepSeek, and gpt-oss locally or in the cloud, with a CLI, REST API, and Python and JavaScript libraries.

ddalcu / mlx-serve

Native macOS and Apple Silicon inference server exposing OpenAI, Anthropic and Ollama APIs, with a Swift app, written in Zig.