Features
Pre-converted .aimodel files for 30+ models
End-to-end conversion scripts with gotcha documentation
Custom Metal kernels for GPU/ANE acceleration
Swift runner (CoreAIKit) for one-line model loading
iOS 27 and macOS 27 Core AI compatibility
Verified on iPhone 17 Pro GPU/ANE
MoE and dense model support
MLA (GLM-4.7-Flash) support
Official-QAT int4 quantization (Gemma 4)
Vision-language models (Qwen3-VL, Holo2-4B, MiniCPM-V 4.6)
Audio understanding (Qwen2.5-Omni-3B) and ASR (Whisper, Qwen3-ASR, Parakeet-TDT)
Text-to-speech (Kokoro-82M, VoxCPM)
Diffusion LLM (LLaDA-8B) and ternary VLA (BitVLA)
Embedding and reranker models for RAG
Object detection models (not explicitly listed but implied by 'detection')
Local model execution for offline inference
Connect online APIs: OpenAI (ChatGPT), Anthropic (Claude), Google (Gemini), Meta (Llama), Mistral AI, Alibaba (Qwen),
123+ HuggingFace models available
Memory feature (coming soon) for context persistence
Minimalist UI with dark mode
Open source under MIT license on GitHub
Cross-platform desktop app (Mac, Windows, Linux)
Version 0.8.3 with bug fixes and performance improvements
