Definition
The Hugging Face Hub is the de facto public repository for open machine learning, hosting over two million models, more than a million datasets, and a comparable number of demo apps (Spaces). For someone building a sovereign AI stack, it plays a role similar to a package registry: it is where you discover and download the open-weight models you then run locally with your own tooling.
Git-based repositories
The Hub stores everything as Git-based, version-controlled repositories, complete with commit history, diffs, and branches. Model repositories carry Model Cards documenting a model's intended use, limitations, biases, and license, which matters when you are choosing a model to depend on rather than a hosted black box. Large files are handled by a chunked storage backend (Xet) to make multi-gigabyte weight downloads practical.
Models, datasets, and Spaces
The "Model Hub" portion specifically indexes models by task, language, and library compatibility, and many are published in formats you can run directly, including GGUF and safetensors. Datasets and Spaces round out the platform, but for self-hosting the key action is the same as pulling open-source software: download the weights, verify the license, and run them on hardware you control.
After downloading a model, run it locally with llama.cpp, Ollama, or vLLM; quantized weights are commonly distributed as GGUF files.
Find open models in the sovereign self-hosting catalog.
In Simple Terms
The Hugging Face Hub is the de facto public repository for open machine learning, hosting over two million models, more than a million datasets, and…