Description
assistant application that runs on touchscreen kiosks. The application should be industry-independent, allowing it to be deployed for real estate, healthcare, hospitality, education, manufacturing, government services, retail, and other sectors simply by changing the knowledge base and media content. The kiosk hardware is already available. We only need the software. Core Features AI Avatar interaction. Voice conversation (Speech-to-Text and Text-to-Speech). Integration with an advanced LLM (OpenAI, Claude, or similar). Knowledge Base powered by uploaded documents. Ability to answer questions based only on uploaded content. Display images, videos, brochures, PDFs, and other media during conversations. Full-screen kiosk interface. Modern and attractive user experience. Admin Dashboard The administrator should be able to: Create new projects. Upload documents (PDF, DOCX, etc.). Upload images and videos. Update or delete content. Rebuild the knowledge base. View conversation history. View basic usage statistics. Technical Requirements Preferred technologies (open to recommendations): React / Next.js Node.js, Python (FastAPI), or similar OpenAI / Claude API Vector Database (Pinecone or equivalent) Speech-to-Text API Text-to-Speech API Avatar API (HeyGen or equivalent) The architecture should be modular, allowing AI providers and third-party services to be replaced in the future without major code changes. Deliverables Complete source code. Admin dashboard. Working kiosk applicati