danielstriver/rwanda-tour-ai

# Rwanda Tourist Assistant Rwanda Tourist Assistant is a RAG-powered (Retrieval-Augmented Generation) AI assistant designed to provide personalized, grounded tourism recommendations for Rwanda. It combines a high-performance React frontend with a serverless backend and a vector database for intelligent discovery. --- ## 🚀 Key Features - **AI-Powered Discovery:** Interactive AI assistant using **Hugging Face (Qwen2.5-72B)** to recommend experiences based on natural language and user preferences. - **Grounded Recommendations (RAG):** AI responses are grounded in a curated knowledge base of over 100+ Rwandan destinations and experiences. - **Advanced Preference Engine:** Filter searches by experience type (Nature, Culture, Adventure), budget (Economy to Luxury), and duration. - **Multilingual Support:** Fully localized in **English, French, and Kinyarwanda**. - **Dark Mode Support:** Modern, immersive UI with seamless theme switching. - **Voice Interface:** Mocked voice assistant with advanced animations and feedback. - **Chat Persistence:** Saved local chat sessions with intelligent title generation. - **API Rate Limiting:** Integrated Upstash Redis to prevent LLM API abuse. --- ## 🛠 Tech Stack - **Frontend:** React 19, TypeScript, Chakra UI, Framer Motion. - **AI/Backend:** Vercel Serverless Functions, Hugging Face Inference API. - **Database (Vector Search):** Supabase Vector (PostgreSQL + pgvector). - **Rate Limiting:** Upstash Redis. - **Deployment:** Vercel. --- ## 📋 Prerequisites To develop on this project, you need the following tools installed: 1. **Node.js & npm:** (v18+ recommended) 2. **Git:** For version control. 3. **Vercel CLI:** For local serverless function development. ```bash npm install -g vercel ``` 4. **Supabase CLI:** (Optional, but recommended for database migrations) ```bash npm install -g supabase ``` 5. **GitHub CLI:** For managing PRs and repos from the terminal. --- ## ⚙️ Setup & Installation ### 1. Clone the Repository ` …