Ollama

Ollama, previously recognized under the names Name TBD and Infra, operates primarily in the B2B sector with a focus on infrastructure. Based in Palo Alto and San Francisco, California, the company was part of the Y-Combinator W21 batch and serves regions across the United States and Canada. Ollama’s platform supports macOS, Linux, and Windows (currently in preview), and is designed to facilitate the local use of large language models (LLMs) including Llama 2, Code Llama, Gemma, Mistral, and LLaVA. These models cater to a variety of tasks ranging from code generation to vision and language understanding, and are available in different parameter sizes such as 7B, 13B, and 34B for vision models. The company offers models with GPU acceleration on Linux and provides an official Docker image for enhanced compatibility. Ollama supports integration with Python and JavaScript libraries, making it accessible for developers to incorporate into their applications. Additionally, the platform includes a search functionality for models, a blog with updates and tutorials, and a GitHub repository for open-source collaboration. Ollama also maintains a Discord community for user engagement and offers guides for integrating LLMs into tools like Obsidian and structuring prompts for optimal model performance.