Mystic Turbo Registry
Mystic Turbo Registry is a custom Docker registry and containerd adapter designed to load ML models significantly faster, up to 11.4x, reducing cold start times by up to 90%. It's a high-performance AI model loader built in Rust to drastically lower cold start time for ML models.
Optimized Docker registry for faster ML model loading and reduced cold starts.
- Reducing cold start times for ML models
- Accelerating container downloading and extraction
- Improving the performance of AI inference
- To use Mystic Turbo Registry
- you can get started with examples like loading an SDXL (10GB Docker image) or Llama-70B (85GB Docker image). The website provides benchmarks and comparisons to demonstrate the speed improvements over traditional methods.
The terminal and Mac menu bar companion for Kickbacks.ai
Screen recordings that edit themselves
See it once. Ask forever.
Keep any Mac window visible in a floating mini screen
Approve AI agent data access with Face ID
Instant, hardware Isolated Sandboxes for AI agents
API integration testing that remembers what breaks
AI-powered schematic design tool for PCB making
AI platform for deep video understanding
Platform for building with Google's Gemini AI models.
Fastest LLM runtime on Apple Silicon
Find & fix live vulnerabilities in Vibe Apps with 1-prompt.
A modern, visual file manager for the terminal
Full-Stack Platform for Training Small Language Models
Tamamon is a free macOS desktop pet that grows as you code with Claude Code. It reads your Claude usage to evolve the pet, but never accesses your code or data. The pet appears on your desktop, reacts to your activity, and can be fed and played with.
Caps Lock keeps your Mac awake, even with the lid closed
Control every aspect of model training and fine-tuning
Scale PostgreSQL without changing your app
