STELLA — Multilingual AI Voice for Hospitality

STELLA is a hospitality-focused AI voice assistant enabling real-time multilingual communication between hotel staff and international guests with natural, instant speech translation.

Comment

Inspiration

In hospitality environments, communication defines the guest experience. Hotels serve international travelers daily, yet language barriers often create misunderstandings, delays, and friction at the front desk.

We were inspired to build STELLA after observing how staff rely on manual translation apps that interrupt natural conversation flow. We asked:

What if communication could be instant, voice-based, and seamless?

That idea became STELLA — a real-time AI voice assistant designed specifically for hospitality.

What it does

STELLA is a real-time multilingual AI voice assistant for hotels.

Staff speak naturally into a device, and STELLA:

Transcribes speech instantly

Translates it contextually for hospitality scenarios

Speaks the translated message aloud in a natural voice

The system supports multiple global languages and is optimized for tablet and front-desk use.

How we built it

We built STELLA as a production-ready MVP using:

Web Speech API for real-time Speech-to-Text

Gemini AI via Lovable edge functions for contextual translation

ElevenLabs API for high-quality Text-to-Speech

Lovable Cloud backend for secure edge function execution

Vercel deployment for hosting

Responsive React-based UI for accessibility and usability

The system processes:

Speech → Text → AI Translation → Natural Voice Playback

All within seconds.

Challenges we ran into

Building a real-time voice workflow introduced several challenges:

Managing latency between transcription, translation, and playback

Handling API rate limits and failures gracefully

Ensuring translations were context-aware (not literal word-for-word)

Securing API keys properly using environment variables

Designing an interface simple enough for non-technical hotel staff

Each challenge helped us refine STELLA into a stable, usable system.

Accomplishments that we're proud of

Built a fully functional, end-to-end real-time voice translation system

Delivered a production-ready MVP rather than a prototype

Implemented secure API key management following least-privilege principles

Designed a clean, accessible UI optimized for hospitality use

Demonstrated applied AI solving a real-world communication problem

What we learned

Through building STELLA, we learned:

Real-time AI systems require thoughtful UX flow design

Latency optimization is critical for voice applications

Context-aware AI is more impactful than literal translation

Security and production readiness matter even in hackathon builds

Applied AI should enhance human interaction, not replace it

What's next for STELLA — Multilingual AI Voice for Hospitality

Our next steps include:

Piloting STELLA in real hospitality environments

Adding intent recognition for common guest requests (check-in, room service, amenities)

Building a conversation insights dashboard for hotel management

Expanding support for more languages and dialects

Integrating with hotel management systems

Releasing a dedicated mobile and tablet application

Our long-term vision is to make STELLA the AI-powered voice layer for global hospitality communication.

Built With

Share this project:

Updates