Libraries and tools for building with AI on Apple silicon.
All Dev tools
Best MLX Score first.
11 of 11
1
TransformersHugging Face's Python framework of model definitions to load, run and train text, vision, audio and multimodal models.★★★★★ 4.7View2
uvA fast Python package and project manager written in Rust; installs Python, tools like mlx-lm and their dependencies.★★★★★ 4.7View3
LiteLLMA Python library and self-hosted gateway that calls 100+ model providers, local or cloud, through one OpenAI-format API.★★★★★ 4.0View4MMLXApple's array framework for machine learning on Apple silicon, the base that MLX models and tools are built on.★★★★★ 3.5View5
Hugging Face CLI (hf)The `hf` command and huggingface_hub library: sign in, download and upload models and datasets on the Hugging Face Hub.★★★★★ 2.7View6CCore ML ToolsApple's Python package for converting PyTorch and TensorFlow models to Core ML, to run them in Mac and iPhone apps.★★★★★ 2.5View7MMLX-AudioText-to-speech, speech-to-text and speech-to-speech on Apple silicon with MLX, as a Python package and command line.★★★★★ 2.5View8LLLMA command line tool and Python library for prompting OpenAI, Claude, Gemini and local models, with plugins and a log.★★★★★ 2.2View9MMLX-VLMRun and fine-tune vision language models on a Mac with MLX, from Python, the command line, a chat UI or a local server.★★★★★ 2.1View10MMLX SwiftThe Swift API for MLX, to build machine learning into Mac, iPhone, iPad and Vision Pro apps with Swift Package Manager.★★★★★ 1.4View11MMLX Swift LMA Swift package for running and fine-tuning language and vision models with MLX in your own Mac and iPhone apps.★★★★★ 1.0View