Ollama: Definition and Examples

Ollama is an open source platform designed to simplify the execution of large language models (LLMs) directly on a local machine. Where using models like Llama, Mistral, or Gemma previously required complex setups, Ollama offers an intuitive command-line interface that automates downloading, configuration, and running models in a few commands.

One of the main advantages of Ollama is data privacy: all interactions with the model remain on the user's machine, with no data passing through external servers. This makes it a particularly popular tool among developers, researchers, and companies concerned about protecting their sensitive data.

Ollama automatically handles model optimization according to available hardware (CPU, GPU, RAM) and provides a local REST API compatible with many tools and frameworks. It supports a large catalog of pre-quantized models, allowing powerful LLMs to run even on consumer hardware without a high-end graphics card.

Thanks to its "Modelfile" system inspired by Dockerfiles, Ollama also allows customizing models by defining system prompts, generation parameters, and conversation templates. This approach makes the tool accessible to both beginners and advanced users who want to create customized AI assistants.