Keeping AI Offline: Ollama and LM Studio

Introduction to Local AI

Running AI locally on devices like the Samsung Galaxy S25 or Apple iPad Air M2 is often seen as a more private alternative to cloud-based services such as ChatGPT. However, some local AI solutions like Ollama and LM Studio may connect to the internet without user intervention, potentially compromising data privacy. This article will guide you through the process of disabling online features in Ollama and LM Studio and ensuring your AI agents remain offline on both Windows and Mac systems.

Understanding the Risks

The Cohere Qwen test revealed that AI agents with shell access can indeed send data out, highlighting the importance of securing your local AI environment. Whether you are using a MacBook Air or a Dell XPS laptop, keeping your AI offline is crucial for maintaining privacy.

Disabling Online Features in Ollama

To ensure Ollama does not connect to the internet, follow these steps on your Windows 10 or macOS device. First, open Ollama and navigate to its settings menu, typically found in the top right corner of the application window. Look for the ‘Network’ or ‘Internet’ settings and toggle off any options that allow the app to connect to the internet. This might include options for automatic updates or data syncing. For users of the Apple Watch Series 10, ensuring that your connected device is secure is also paramount.

Securing LM Studio

LM Studio, another popular local AI solution, also requires careful configuration to keep it offline. Upon launching LM Studio on your computer, whether it’s a desktop like the HP Envy or a laptop such as the Lenovo ThinkPad, go to the application’s preferences or settings. There, you should find options related to online connectivity. Disable any features that could potentially allow LM Studio to send or receive data over the internet. This step is crucial for protecting your privacy, especially if you are working with sensitive information on devices like the iPad Pro.