I need a small, public-facing website whose sole purpose is to chat with visitors about Prayagraj and guide them to current cultural events—festivals, exhibitions, live performances, campus happenings, anything that gives the city its character. Here is what I have in mind: the landing page loads a conversational widget powered by an AI model (OpenAI, Hugging Face, or a comparable LLM—whichever you are most comfortable using). When a traveller asks “What’s on this weekend?” or “Is there a classical music recital tonight?”, the assistant responds with up-to-date, location-specific tips plus practical details such as venue, timings, ticket links and a quick route description. It must recognise follow-up questions, keep context for at least a few turns, and politely nudge the user if the query drifts outside Prayagraj. Key goals • Clean, mobile-first UI (React, Vue or vanilla JS—your call) • Simple CMS or Google-Sheet bridge so I can refresh event data without touching code • Clear documentation: set-up steps, environment variables, how to retrain or fine-tune the model for new event categories • Deployment to a low-cost host (Vercel, Netlify, or similar) with HTTPS enabled Acceptance criteria 1. User can ask at least five varied cultural questions and receive accurate, distinct answers. 2. “Near-real-time” event list must update from the data source in under 10 minutes after I edit it. 3. All third-party APIs and keys are hidden from the client bundle. 4. Page loads in under three seconds on 4G according to Lighthouse. Hand-over deliverables • Source code repository • Brief video or screenshots demonstrating the assistant in action • One-page technical walkthrough If something here needs clarification before you start, let me know so we can pin it down quickly.