Run models privately

Run inference on your own hardware, with room for the whole codebase and none of your data leaving the building.