Description of problem:

The Developer Lightspeed configuration procedure had several UX issues that made it confusing for users, especially those starting from scratch:

Issues Identified

1. Unclear "Bring Your Own Model" Setup
   •  The `start-lightspeed.sh` script offers a "bring your own model" option, but the README.md didn't clearly explain how to configure it
   • Users were unsure whether to use VLLM variables, OpenAI, or other providers
   • The connection between script choices and required environment variables was not obvious

1. Container Image Configuration Duplication
   •  Container images for Llama Stack and Lightspeed server were set in two places:
     •  `default.env` file
     • `developer-lightspeed/compose.yaml` defaults
   •  The images in these locations didn't match, causing confusion
   • Users didn't know which source of truth to follow
2. Insufficient Documentation Comments
   • Configuration files lacked explanatory comments
   • Users couldn't understand what each variable controlled or when to use it
3. No Stop Script
   • Only a start script existed, making it difficult to restart or cleanup
   •  Users had to manually remember which compose files to use for cleanup

Prerequisites (if any, like setup, operators/versions):

Podman/Docker

Steps to Reproduce

 # Use rhdh-local, and follow the developer-lightspeed documentation to setup lightspeed.

Actual results:

Confusing UX for the users who are just getting started.

Expected results:

Improve UX for the new users.

Reproducibility (Always/Intermittent/Only Once):

Always{}