Skip to main content
The predict command runs a compiled model directly from your terminal. This is useful for quick testing and scripting.

Usage

Options

Running Inference

Pass inputs to your compiled model using --param value syntax:
For models with multiple parameters:

Quiet Mode

Use --quiet to suppress logging output, which is useful for scripting:
The predict command outputs results to stdout, making it easy to pipe into other commands or capture in scripts.

Exploring Models

Not sure what models are available? Use the --explore flag to browse:
For production use cases, consider using the Muna SDK in your preferred language for better error handling and performance.

Next Steps

Using the Muna Client

Learn how to run inference with the SDK.

Explore Models

Browse public models on Muna.