You can run powerful AI models in-house directly on your PC or server using free, open-source tools. These options work offline, protect privacy, and avoid recurring subscription costs. They are built to run on Windows, Mac and Linux.
As a first step go to https://www.canirun.ai/ to assess your system’s ability to run various models. Enter data including GPU, Processor and Ram to see which models can be supported on your hardware.
For piloting any model, you’ll want a minimum of 8GB Ram and a modern processor. For larger models, you’ll need a minimum of 16GB Ram, a GPU and SSD storage. For production work, you’ll want to invest in robust hardware.
Local, open-source models you can download and install include: Local ai, Ollama, LM Studio, GPT4All, and Jan.
Here’s a video with steps to install Ollama on Windows
Want to pilot your own internal AI engine? Our experts can help make it happen. Contact us!