Osaurus brings both local and cloud AI models to your Mac

Osaurus has announced the release of a new Mac application that enables users to run and switch between local and cloud AI models within a single interface, marking a significant step toward more private and customizable AI experiences on personal hardware.

The Osaurus app is an open-source, Apple-only platform that connects users to a variety of AI models, both locally hosted and cloud-based, including models from OpenAI, Anthropic, and Apple’s on-device models. It allows users to manage and switch between models easily, with security measures such as hardware-isolated sandboxes to protect data.

The software supports models like MiniMax M2.5, Gemma 4, Qwen3.6, GPT-OSS, Llama, and DeepSeek V4, among others, and can connect to cloud providers including OpenAI, Anthropic, and xAI. It also features over 20 native plugins for tasks like email, browsing, and file management. Recently, voice capabilities were added to enhance user interaction.

Developed by Osaurus’s team, including co-founder Terence Pae, the app aims to address the resource-intensive nature of local AI by advocating for future hardware improvements. Pae highlighted the growing sophistication of local AI, noting that models now perform complex tasks like coding, browsing, and e-commerce orders, with hardware requirements currently high but expected to decrease over time.

Why It Matters

This development is significant because it offers a more private, secure, and customizable alternative to cloud-based AI services, appealing to individual users and organizations concerned about data privacy. By enabling local AI processing, Osaurus could reduce reliance on data centers, potentially lowering energy consumption and operational costs. It also positions Mac users at the forefront of personal AI deployment, with implications for enterprise use in sensitive sectors like healthcare and legal services.

Express Schedule Free Employee Scheduling Software [PC/Mac Download]

• User-friendly drag & drop interface: Simple shift planning
• Manage time-off and leave: Add sick leave, breaks, holidays
• Email schedules to employees: Direct email scheduling

View Latest Price

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Background

Over the past year, the AI community has seen increased interest in local models, driven by privacy concerns and hardware advancements. Existing tools like OpenClaw and Hermes provided some control over AI models but often targeted developers and posed security risks. Osaurus’s approach simplifies this for consumers, integrating multiple models and plugins into a user-friendly interface. The move aligns with broader industry trends toward decentralizing AI deployment and reducing cloud dependency.

“The goal was to run an AI assistant locally, giving users full control over their data and tools while accessing powerful models directly on their Mac.”

— Terence Pae, Osaurus co-founder

“Local AI’s needs will come down over time as hardware improves. We’re seeing models now capable of complex tasks, and this will only get better.”

— Pae

What Remains Unclear

It remains unclear how widely adopted Osaurus will become among consumers and enterprises, especially given current hardware requirements. The scalability of running larger models on typical Macs and the performance in real-world scenarios are still untested at scale. Additionally, the long-term security implications of running AI models locally versus in the cloud are still being evaluated.

What’s Next

Next steps include expanding compatibility with more AI models and plugins, improving performance and hardware efficiency, and potentially offering enterprise-focused versions. Osaurus’s team is also participating in startup accelerators, hinting at future funding or commercialization efforts. Monitoring user adoption and real-world performance will be key milestones.

Key Questions

Can I run Osaurus on my existing Mac?

Yes, but running larger models may require Macs with at least 64 GB of RAM, and more powerful systems for bigger models like DeepSeek v4.

Does Osaurus support all AI models I want?

It supports a wide range of models, including popular open-source and proprietary options, with ongoing updates to expand compatibility.

Is Osaurus secure for sensitive data?

Yes, it runs models in hardware-isolated sandboxes, which helps protect user data and privacy.

Will this reduce my dependence on cloud AI services?

Potentially, as local models become more capable and hardware requirements decrease, users may rely less on cloud providers for AI tasks.