Skip to main content



APIs to run inference with MAX Engine.

Although there are several modules in this max.engine package, you'll get everything you need from this top-level engine namespace, so you don't need to import each module.

For example, the basic code you need to run an inference looks like this:

from max import engine

def main():
# Load your model:
var session = engine.InferenceSession()
var model = session.load(model_path)

# Get the inputs, then run an inference:
var outputs = model.execute(inputs)
# Process the outputs here.

For details, see how to run inference with Mojo.


  • info: Provides information about MAX Engine, such as the version.
  • model: Defines the Model type that holds a model ready for execution.
  • session: Defines the InferenceSession type that serves as an entry point to MAX Engine.
  • shape_element: ShapeElement API. See ShapeElement struct definition for details.
  • tensor: Defines different data formats you can use to pass inputs to MAX Engine when executing a model.
  • tensor_map: Defines the TensorMap type that holds input and output tensors for a model.
  • tensor_spec: Defines EngineTensorSpec to define a model's input and output specs.
  • value: Defines different value types you can be pass in and out of models.