Osaurus brings both local and cloud AI models to your Mac
A Unified AI Experience for Mac Users
Osaurus is a unique open source platform designed exclusively for Mac, offering seamless integration of both on-device and cloud-based AI models. The project started as an evolution of Dinoki, an AI-powered desktop assistant, after early users expressed interest in running AI locally to avoid ongoing token fees and maintain data privacy.
Flexible Integration with Local and Cloud Models
Osaurus lets users toggle between various AI models—either running them directly on their Mac or connecting to popular cloud services like OpenAI and Anthropic. Your files, tools, and model memories all stay secured on your hardware, enhancing both privacy and convenience.
Easy-to-Use Interface with Robust Security
Unlike many developer-focused AI tools, Osaurus features a hardware-isolated sandbox and a consumer-friendly interface, minimizing security risks and delivering a safe and simple experience for everyday users.
Model and Plug-In Support
- Compatible with leading models: MiniMax M2.5, Gemma 4, Qwen3.6, GPT-OSS, Llama, DeepSeek V4, and more.
- Supports Apple’s and Liquid AI’s on-device models.
- Connects to cloud providers such as OpenAI, Gemini, xAI/Grok, OpenRouter, and Ollama.
- Works as a Model Context Protocol (MCP) server and includes over 20 native plug-ins (Mail, Calendar, Vision, Browser, Filesystem, etc.).
Recently, voice capabilities were also added to enhance user interactions.
Hardware Requirements and Forward Outlook
Running advanced models locally requires a Mac with significant RAM (64GB for general use, 128GB for larger models like DeepSeek v4). However, co-founder Terence Pae is optimistic that improvements in AI efficiency will soon lower these hardware demands.
Positioning Osaurus in the Market
With over 112,000 downloads, Osaurus stands out among tools like Ollama and LM Studio by prioritizing usability for non-developers. The team is exploring business applications, especially in fields where data privacy is critical, such as legal and healthcare sectors. There's also a vision for on-premises deployment to reduce reliance on centralized AI data centers.
For more details, see the original article at TechCrunch.
