ramalama/docs/demo

RamaLama

This script demonstrates running RamaLama with a sample workflow that pulls a model, serves it, and allows testing inference through a browser or curl.

Requirements

• RamaLama installed and available in your PATH
• Podman installed and configured

Usage

Run the script:

./ramalama.sh

Override the browser (optional):

BROWSER=google-chrome ./ramalama.sh

Features

• Pulls and runs the smollm:135m and granite models with RamaLama
• Opens the service endpoint in your browser automatically
• Waits for the service to be ready before testing inference
• Performs a sample inference with curl against the granite3.1-dense model

Advanced usage

You can also call specific functions from the script directly, for example:

./ramalama.sh pull
./ramalama.sh run
./ramalama.sh test

Extra arguments can be passed after the function name if supported.

---