LM Studio Apple MLX engine

High-performance MLX-based LLM inference engine for macOS with native Swift implementation

Machine Learning eXchange (MLX). Data and AI Assets Catalog and Execution Engine

AI-Powered Autonomous Binary Reverse Engineering CLI — the native reverse engine from Innora-Sentinel. Local LLM inference (MLX), MPS GPU acceleration, multi-round iterative analysis, zero API cost.

Local Video RAG Engine. A FastAPI microservice for video understanding: Scene Detection + Whisper ASR + Qwen3-VL. Optimized for Apple Silicon (MLX) & Windows/Linux (Llama.cpp).

Local LLM benchmark tool for comparing engines (MLX vs llama.cpp), scenarios on Apple Silicon

mlx port of Karpathy's micrograd - a tiny scalar-valued autograd engine with a small PyTorch-like neural network library on top.

Parakeet MLX is a next-generation automatic speech recognition (ASR) engine optimized for Apple Silicon (M1/M2/M3), leveraging Apple’s MLX framework for ultra-fast, low-latency transcription. It offers real-time streaming, advanced audio processing. Including noise reduction and silence detection

LoRA fine-tuning gradients on Apple Neural Engine via private APIs. Conv-as-matmul, spatial constraints, MLX integration.

A python training pipeline for Apple MLX to produce a fine tuned model with Apple Neural Engine Support

High-performance inference engine — Rust/MLX/GGUF

vdub — video dubbing and subtitle engine. Rust + MLX. Free local ASR/TTS.

A Cub3D raycasting project parser and basic renderer built in C using the MiniLibX (mlx) graphics library. This project parses .cub map files, validates map data and textures, and sets up the data structures required for a raycasting engine similar to the classic Wolfenstein 3D style.

A Cub3D raycasting project parser and basic renderer built in C using the MiniLibX (mlx) graphics library. This project parses .cub map files, validates map data and textures, and sets up the data structures required for a raycasting engine similar to the classic Wolfenstein 3D style.

Native Apple Silicon LLM inference engine powered by MLX. Single binary, zero Python. 124+ tok/s on Qwen3-4B-4bit.

BitNet 1-bit LLM inference on Apple Silicon: 89 tok/s with custom Metal kernel. Includes CoreML/ANE benchmarks and speculative decoding.

No description available

A lightweight autograd engine built using numpy and MLX (<1000 LoC)

High-performance inference engine for large language models optimized for Apple Silicon (MLX backend)

A 3D raycasting engine inspired by Wolfenstein 3D, implemented in C using the MLX graphics library.

AlphaZero‑inspired chess engine for Apple Silicon, combining an MLX neural network, C++ MCTS UCI engine, and Swift orchestrator for self‑play training, evaluation, and benchmarking.

Rust MLX inference engine for Qwen3.5-35B-A3B on Apple Silicon — 112 tok/s single-user, 200+ tok/s multi-user

Cross-modal vision-language search engine. Index text and images, search across modalities. Multi-backend (PyTorch/MLX), tiered modes, MCP server. Inspired by QMD by Tobi.

No description available

MLX compute engine for Apple Silicon

No description available

MLX model inference engine

No description available

miniRT 개발을 위한 engine 개발 레포 by @JuneParkCode @kimminkyeu

Context Engineering Framework for MLX Trading Pipeline - Comprehensive AI development context with Apple Silicon optimization patterns, PRP templates, and implementation guidelines