raullenchai
To provide the fastest local AI engine for Apple Silicon, enabling high-performance LLM inference with native tool calling and prompt caching.
At a Glance
- Apple Silicon users
- Open source developers
- AI researchers
- Privacy-conscious AI users
AI Tools by raullenchai
(1)Rapid-MLX
Local AI Inference for Apple Silicon
Discussions
No discussions yet
Be the first to start a discussion about raullenchai
Latest News
Rapid-MLX v0.6.15 released with post-v0.6.14 batch and codex review fixes.
Introduction of the single-command merge-readiness pipeline for automated PR grading.
Launch of the Model-Harness Index (MHI) and Homebrew distribution support.
Rapid-MLX adopts Apache 2.0 license and formalizes community contribution templates.
Products & Services
A high-performance local AI inference engine optimized for Apple Silicon, featuring 17 tool parsers, prompt caching, and reasoning separation.
Market Position
2-4x faster than Ollama and llama.cpp on Apple Silicon. Only engine supporting day-0 features for MoE models like DeepSeek V4 Flash on Mac.
Leadership
Founders
Raullen Chai
Ph.D. in Encryption from University of Waterloo. Former Lead of Crypto R&D and Engineering Security at Uber. Former engineer at Google and Oracle. Co-founder and CEO of IoTeX/MachineFi Lab.
Executive Team
Raullen Chai
Founder and Lead Developer
Cybersecurity and cryptography expert, CEO of IoTeX.
Founding Story
Started to address the performance gap of existing local LLM engines like Ollama on Apple Silicon by leveraging the MLX framework for native Metal compute kernels.
Business Model
Revenue Model
Open-source project (Apache 2.0). No formal revenue model disclosed for the software itself.
Pricing Tiers
Available on GitHub and PyPI.
Target Markets
- Apple Silicon users
- Open source developers
- AI researchers
- Privacy-conscious AI users
- Local LLM serving for developers
- AI agents (Cursor, Claude Code, Aider)
- Private AI infrastructure
- Apple Silicon performance benchmarking
- Cursor
- Claude Code
- Aider
- PydanticAI