fix: swap decommissioned llama-3.3-70b-versatile for openai/gpt-oss-120b

fix: correct local NEXT_PUBLIC_API_URL example to port 8000

docs: update deploy references from Vercel to Netlify

ui: add favicon; trim footer to model line

prep for hf deploy

style: replace box diagram with a list, normalize unicode to ASCII, trim comments

feat: Next.js frontend for hosted demo + README live-demo section