Hugging Face
Also known as: HF, HuggingFace, Hugging Face Hub
Hugging Face is where most open-source AI models live. The Hub hosts hundreds of thousands of model checkpoints — from small embedding models to massive frontier-scale open-weight LLMs — alongside the datasets used to train them and demo apps (called Spaces) that let you try models in a browser. If you want to use a model that is not from OpenAI or Anthropic, your first stop is almost always Hugging Face.
For builders, Hugging Face is useful at multiple stages. During evaluation, the free Serverless Inference API lets you test thousands of models without standing up any infrastructure. For production, Inference Endpoints let you deploy a specific model on dedicated hardware managed by Hugging Face. The Hub also acts as a version-controlled registry for model weights (the saved numerical parameters of a model), so teams can track exactly which checkpoint is in production.
Hugging Face also maintains popular Python libraries like Transformers and Diffusers that standardize how models are loaded and run. Many AI frameworks and serving tools plug directly into the Hub, making it easy to go from browsing a model page to running that model in your app.