fix: clear input fields when clearing history

design: complete UI redesign — remove all emojis, premium typography, refined layout, professional aesthetic

fix: robust decision extraction (handle missing closing tag) and always show a result

fix: force moderator to always produce DÉCISION FINALE on first and only round

feat: enforce extreme conciseness and reduce max rounds to 2 to save tokens and speed up consensus

feat: add conversational continuation and clear history features

fix: reduce max_tokens and add retries to handle Groq TPM rate limits gracefully

fix: switch default model to llama-3.1-8b-instant to bypass 70b daily quota limit

fix: reduce stop sequences to 4 to comply with Groq API limits

fix: bypass ZeroGPU quota consumption with dummy function

fix: bind stop sequences to llm to prevent hallucination loop and rate limit 429 errors

fix: enforce template translation and remove hardcoded UI titles

fix: enforce language detection with explicit SystemMessage and improve final decision structure

fix: prompt engineering for strict multilingual support, anti-hallucination, and better debate contradiction

feat: add automatic multilingual support and optional location field for restaurant recommendations

style: remove emojis and update agent names to be more professional

fix: add @spaces.GPU decorator for ZeroGPU compatibility

fix: disable SSR mode for free-tier HF Spaces compatibility

fix: remove show_copy_button (deprecated in Gradio 6.x)

fix: Gradio 6.x compatibility — move css/theme to launch(), remove deprecated Chatbot params

feat: implement Madia multi-agent debate app

chore: add project structure with .gitignore

initial commit