Travel Culture and Food AI mood

The project provides a full-stack assistant intended for e-commerce or travel scenarios. The backend is written in FastAPI and exposes several routes. The /agent endpoint relies on language model and search API keys; missing credentials cause a runtime error surfaced to the frontend. The /speech route accepts audio, returning a transcription and a base‑64 MP3 audio URL, with optional translation to English. Another /notify endpoint allows sending SMS messages through Twilio. Additional routes include search and map services implemented in the backend module. The frontend, built with Next.js, communicates with these endpoints and offers interactive components such as a chat box, voice recorder, and a map view. Users can run npm run dev to start the frontend on http://localhost:3000. Suggested results appear as buttons that automatically send follow‑up prompts when clicked, extracting list items from the language-model response if present. A “Color Settings” menu lets users switch themes or set a custom background image. Overall, the application combines a FastAPI backend (handling LLM queries, speech processing, searches, map coordinates, and SMS notifications) with a Next.js frontend that supports interactive chat, audio input, map display, and customizable visual themes.