Llama-Stack Playground

Goal

This module introduces the Llama-Stack Playground, a user-friendly interface designed to demonstrate the core capabilities of Llama Stack in an interactive environment. The dashboard allows you to inspect API providers and resources exposed by Llama Stack.

Prerequisites

Before starting, ensure you have the following:

  • A running Llama Stack server (see: Llama-Stack Hello World)

  • Python 3.10 or newer (to install and run the CLI tools)

Step 1: Deploy the Llama-Stack Playground

You can deploy the playground using the source code or run the pre-built container image from Quay.io:

podman run -p 8501:8501 \
  -e LLAMA_STACK_ENDPOINT=http://host.containers.internal:8321 \
  quay.io/rh-aiservices-bu/llama-stack-playground:0.2.1

Once the container is running, the Llama-Stack playground will be available at http://localhost:8501.

Step 2: Interact with the Playground Chat

The Llama-Stack Playground is built with Streamlit and includes a chat interface for interacting with the Llama Stack server.

Open your browser and navigate to http://localhost:8501. You’ll see a chat interface where you can enter prompts and receive responses from the deployed LLM model.

Llama-Stack Dashboard

Step 3: Explore the Playground

The playground also allows you to explore available resources, such as models, vector databases, and shields.

To browse these:

  • Click the "Resources" tab to view all registered components in your Llama Stack setup.

Resources

  • Click the "API Providers" tab to inspect configured model and embedding providers.

API Providers

Summary

In this module, you:

  • Deployed and launched the Llama-Stack Playground

  • Interacted with the LLM model through the chat interface

  • Explored available models, providers, and other resources

Next, continue with Llama-Stack Python Programming to write and run your first Python program using Llama Stack.