Skip to content
Mani Paul
FeaturedActive

Nurae

Multimodal AI Voice Agent Platform

Nuronics Labs · 2025 – Present · Senior Software Engineer – Application Development

Overview

Nurae is a next-generation multimodal AI platform that enables users to accomplish real-world tasks through natural conversations. Instead of navigating multiple applications or websites, users can simply interact with an AI assistant using voice, text, or a phone call. The platform is designed to orchestrate intelligent workflows that connect users with services such as transportation, travel planning, local businesses, and other digital experiences.

Problem

Most digital services require users to install dedicated applications, learn different interfaces, and manually complete every step of a task. This creates unnecessary friction, especially when users simply want to accomplish an outcome rather than interact with multiple applications.

Solution

Nurae combines real-time conversational AI, multimodal interactions, modern backend services, and mobile technologies into a unified platform. Users can interact through mobile, voice, text, or direct phone calls via Twilio. The architecture is provider-agnostic, allowing multiple AI models and services to be integrated as the platform evolves.

Architecture

  1. 1React Native (Expo)
  2. 2Next.js Web Dashboard
  3. 3FastAPI Backend
  4. 4LiveKit Agents
  5. 5Multiple AI Providers
  6. 6Supabase

React Native (Expo) → Next.js Web Dashboard → FastAPI Backend → LiveKit Agents → Multiple AI Providers → Supabase

Responsibilities

  • Developed backend services using FastAPI.
  • Built the React Native mobile application.
  • Integrated LiveKit Agents for real-time AI conversations.
  • Designed APIs supporting multimodal interactions.
  • Implemented authentication and user management.
  • Contributed to system architecture and production-ready application design.

Challenges

  • Building reliable real-time conversational experiences with minimal latency.
  • Designing a flexible architecture capable of supporting multiple AI providers.
  • Maintaining consistent experiences across voice, text, and mobile interactions.
  • Supporting direct phone-call interactions using Twilio without requiring app install.

Tech stack

Backend

FastAPIPython

Frontend

React Native (Expo)Next.jsTypeScript

Infrastructure

SupabaseDocker

AI & Media

LiveKit AgentsMultiple AI ProvidersTwilio

Screenshots

Nurae mobile app interface
Nurae mobile application
Nurae web dashboard
Nurae web experience

Impact

Nurae is currently under active development as a production AI platform focused on delivering natural conversational experiences across voice, text, mobile applications, and telephony.

Lessons learned

  • Designing conversational AI platforms requires balancing flexibility, scalability, and user experience from the beginning.
  • Supporting multiple AI providers reduces vendor lock-in and enables experimentation with different models.
  • Building modular services early simplifies adding new communication channels and integrations.
  • Real-time voice applications require careful consideration of latency, reliability, and conversational flow.

Future improvements

  • Multi-agent orchestration
  • Expanded MCP integrations
  • RAG-powered knowledge retrieval
  • Enterprise administration dashboard
  • Workflow automation
  • Additional communication channels