Ollama builds a platform for running open-source large language models on local machines. Its open-source runtime lets developers execute models such as Llama 3, Gemma, Qwen, GLM, MiniMax and DeepSeek with a single command, on macOS, Windows and Linux. When models run locally, prompts do not leave the device and are not stored or used for training. For workloads requiring more compute, the company offers a cloud-hosted model service that remains compatible with existing tools through OpenAI and Anthropic API support.
The engineering work spans local model inference, MLX optimisation for Apple Silicon, NVIDIA and AMD GPU acceleration, cloud-hosted inference and developer tooling. The platform has been adopted by over 8.9 million monthly active developers, with more than 1 billion model downloads, over 200 trillion tokens processed and over 67,000 community-built integrations.
Founded in 2023 as an open-source project, the company operates in the developer tools and open-source AI sectors. It has raised $88 million in funding. Its stated priorities include developer control, privacy, cross-platform support and keeping the process of running open models simple.
