[Suggestion] Expo React Native real-time voice agent with Deepgram Agent SDK

[deepgram-robot]

What to build

A React Native (Expo) mobile app that connects to Deepgram's Voice Agent API for real-time conversational voice interactions, with a polished mobile UI including animated voice visualization and conversation history.

Why this matters

Mobile voice agents are the fastest-growing segment in voice AI, yet Deepgram has zero mobile examples that go beyond basic transcription. Developers building voice-enabled mobile apps — customer service bots, health assistants, language tutors — need a reference showing how to handle microphone permissions, audio streaming, and agent conversation state on mobile. The existing React Native example covers only basic STT; this covers the full voice agent experience.

Suggested scope

• Language: TypeScript
• Framework: Expo SDK 52+, React Native
• Deepgram APIs: Voice Agent API (WebSocket)
• Features: Microphone capture with expo-av, real-time audio streaming, agent response playback, conversation transcript display, animated voice indicator
• Complexity: Medium-high — requires mobile audio handling, WebSocket management, and UI polish
• Expo managed workflow (no native modules required)

Acceptance criteria

• Runnable with minimal setup (clone, add API key, npx expo start)
• README explains the pattern clearly
• Uses current SDK version
• Works on both iOS and Android via Expo Go
• Handles microphone permissions gracefully
• Shows real-time conversation transcript alongside voice interaction
• Includes voice activity indicator animation

---

Raised by the DX intelligence system.