d-hackmt/LLM-GATEWAYS ? reverse-engineered prompt

Build me a simple demo notebook that shows how an LLM gateway works with Portkey, using a few realistic Python examples and clear explanations for a beginner.

I want it to show how one app can send the same kind of LLM request through Portkey instead of calling a provider directly, then demonstrate the main benefits, like routing, retries, timeouts, fallback providers, load balancing, caching, and request metadata for tracking cost and usage. Please make it easy to follow, with short code cells, plain language notes, and a couple of example requests that would work with providers like Groq or OpenAI. If you need current setup details from the docs, look them up online.

It would also be great if the notebook explains the difference between a provider slug and a config ID, since that seems important here. Keep it practical and polished, so someone can run through it and understand why a gateway helps without changing their app logic.