AI Tour Guide Generator – 3D Building Tour Builder
Full-stack AI tour builder with panorama viewing, LLM image narration, TTS, and Supabase persistence
Built a deployed full-stack application for creating AI-guided 3D building tours. The system lets users create tour projects, upload room photos or panorama images, select LLM providers, generate scene descriptions, edit guide scripts, render Vietnamese female narration, and open an immersive fullscreen tour player with audio and floating room information.
Timeline
2026
Type
Project
Status
completed
My work
- •Built the tour builder UX, FastAPI endpoints, LLM provider abstraction, TTS endpoint, and persistence flow
- •Deployed the Next.js frontend and added demo fallbacks for local or no-key usage
Outcome / Impact
- •Shipped a deployed frontend for building and playing AI-generated building tours with panorama and normal-image support
- •Implemented multi-provider LLM runtime for image description and tour-script generation with Gemini, Groq, OpenAI, Claude, and mock fallback
- •Built project persistence with Supabase plus localStorage/artifact fallbacks for demos without cloud configuration
- •Added Vietnamese TTS generation, reusable audio storage, browser speech fallback, drag-and-drop uploads, URL image ingestion, and sample panoramas
Tech / Skills
Project Media
Demo video and visual walkthrough for this project.
Project Screenshots
Case Study
1) Context / Problem
Real-estate, campus, and facility tours often require manual scripting, separate media handling, and disconnected narration workflows. The project explored how LLM vision, text editing, TTS, and panorama viewing can be combined into a practical tool for producing guided building tours quickly.
2) Your Role
I built the application across frontend, backend, AI integration, persistence, and demo assets. The work covered the tour builder UX, FastAPI endpoints, LLM provider abstraction, TTS endpoint, Supabase storage/database integration, and deployed Next.js frontend.
3) Approach
Structured the system around a Next.js tour builder and FastAPI service layer. Users can upload images or provide image URLs, request AI descriptions through a provider-agnostic LLM interface, edit generated scripts, render narration through VieNeu-TTS, and play the result in a fullscreen tour experience with panorama support and fallback modes.
4) Result / Impact
Delivered a deployable AI tour product with landing page, builder, player, audio guide, background music, sample artifacts, Supabase persistence, and local fallback behavior. The app can run with mock AI for demos or connect to Gemini, Groq, OpenAI, and Claude when API keys are provided.
5) Learnings
AI product reliability depends on graceful fallbacks as much as model quality. Provider abstraction, local demo modes, reusable generated assets, and clear error/loading states made the product usable even when external AI or storage services were unavailable.
6) Links
See links above.