Where developers grow together

AgoraCare - Voice AI Healthcare Companion for Rural India

Voice AI Hackathon 2026 · AI Mobile Coders × Agora · Submission A real-time Hindi/English Voice AI healthcare companion for rural India. Patients tap one button, speak in Hindi, and talk to an AI doctor — get medication reminders, log vitals by voice, and get instantly connected to a live human nurse over an Agora audio bridge in emergencies. No typing. No English required. Built for ₹5,000 Android phones. The Problem 1 doctor per 1,456 people in rural India (WHO standard: 1 per 1,000) Elderly patients forget medications — non-adherence is a major cause of preventable harm Emergency response is slow — patients don't know when or how to escalate Language barrier — most health apps are English-only and text-first AgoraCare solves this with voice. A 68-year-old farmer who can't read English can tap one button and speak. Key Features Feature What it does 🎙️ One-tap Voice Orb Tap to talk to "Aria", the AI health companion, in Hindi or English. Real-time transcript below the orb. 💊 Medication management Today's schedule with status chips (Upcoming / Due Now / Taken / Missed) and an adherence ring 📊 Vitals tracking Blood pressure and heart-rate logging with 7-day trend charts 🚨 Emergency escalation One tap connects the patient to a live nurse via Agora RTC audio bridge; the nurse dashboard gets a real-time ticket with alert sound 👩‍⚕️ Nurse dashboard Real-time escalation tickets from Firestore — join the patient's audio channel or resolve 🌐 Hindi / English Full bilingual voice experience, switchable live Why Agora Conversational AI Is the Core Voice isn't a feature in AgoraCare — it is the entire user experience. Every interaction flows through Agora: agora_rtc_engine carries all real-time audio (patient ↔ AI agent "Aria") The backend starts an Agora Conversational AI agent in the patient's channel with their language + medical context Emergency escalation bridges a live nurse into the same Agora channel — human-in-the-loop, in real time Architecture Flutter Android App ├── Voice Orb ──▶ Agora RTC Engine (agora_rtc_engine ^6.3.2) │ ↕ audio stream │ Agora Conversational AI Agent "Aria" (Hindi + English) │ ↕ tool calls ├── Next.js Backend (Vercel) │ /api/agora/agent/start · /api/agora/agent/stop · /api/agora/token │ ↕ └── Firebase (Auth · Firestore) users · medications · adherence · vitals · escalations (real-time) Screens: Splash → Onboarding (Hindi/English) → Anonymous sign-in → Home (4 tabs: Voice · Medications · Vitals · Profile) → Nurse Dashboard (real-time escalation view) Tech Stack Mobile: Flutter 3 (Dart), Riverpod state management Voice AI: Agora RTC Engine ^6.3.2 + Agora Conversational AI Backend: Next.js on Vercel (token server + agent lifecycle + emergency webhook) Data: Firebase Auth (anonymous sign-in), Cloud Firestore (real-time listeners) UI: Material 3 light theme, Inter font, custom waveform painter, fl_chart Getting Started Prerequisites Flutter SDK >=3.0.0 <4.0.0 An Agora project (App ID + App Certificate) with Conversational AI enabled Firebase project with Android app registered (google-services.json) The AgoraCare backend deployed (Next.js) — see AppConstants.backendUrl in lib/core/constants.dart Run flutter pub get flutter run Configure lib/core/constants.dart — set your Agora App ID and backend URL android/app/google-services.json — add your Firebase config flutter build apk --release (use --split-per-abi to keep the APK lean) Android permissions required: RECORD_AUDIO, INTERNET, BLUETOOTH, MODIFY_AUDIO_SETTINGS. Project Structure lib/ ├── main.dart · app.dart ├── core/ # constants, theme (design system), routes ├── services/ # agora_service, backend_service, firebase_service ├── providers/ # voice, medication, auth (Riverpod) ├── screens/ # splash, onboarding, auth, home, voice, │ # medications, vitals, profile, nurse dashboard └── widgets/ # voice orb, waveform painter, cards, charts Full product spec: PRD.md Judging Criteria — How AgoraCare Maps Criterion Our answer Innovation Human-in-the-loop nurse escalation via a live Agora RTC audio bridge Technical Implementation Native Flutter + agora_rtc_engine — reliable on low-end Android, no browser quirks Effective Use of Agora Conversational AI Voice AI is the whole UX, not a feature — every flow runs through it User Experience One-tap voice orb, 56px touch targets, large touch targets, clean light theme, Hindi-first voice Real-World Impact Built for the 65% of India in rural areas with limited healthcare access Team Built by Amaan Ahmad (@iamaanahmad) - for the Voice AI Hackathon 2026 (AI Mobile Coders × Agora). License MIT — see LICENSE (to be added).

VoiceTwin — AI Communication Twin

VoiceTwin - AI Communication Twin is a real-time AI communication studio powered by Agora Conversation AI and AI Agent, designed to help people practice the conversations that matter. Instead of simply chatting with an AI, VoiceTwin creates realistic voice-based scenarios for interviews, presentations, meetings, client pitches, networking, and high-pressure conversations. Its Communication Twin understands the user's goals, context, communication preferences, and practice history to make each session more personalized. Users can interact with adaptive AI personas, handle real-time follow-up questions, enter 30-second Pressure Mode, and receive detailed insights on clarity, confidence, relevance, conciseness, technical depth, pacing, and speech patterns. VoiceTwin also includes 60s Pitch, STAR Reframer, Speech Lab, Insights Report, Twin DNA, Progress Curve, and personalized practice recommendations to help users identify communication gaps and continuously improve. The core idea is simple: VOICE → CONVERSATION → ADAPTATION → INSIGHT → IMPROVEMENT VoiceTwin turns AI voice interaction into a personalized communication practice studio — helping users speak better, think faster, and communicate with confidence.

Roundtable AI - One Candidate. Three AI Interviewers.

Roundtable AI turns a job description / resume into a live voice interview. Interview practice today is either a generic question bank or a single chatbot that asks whatever it feels like. Neither prepares you for the role you are actually applying to. You upload the job description, or your resume. Roundtable AI parses it into a structured profile of the role: required skills, system design areas, responsibilities. Three specialist interviewers panel then question you over live voice, one round each: technical, system design, behavioral. Each has a different voice and persona. When it ends you get a scored evaluation with evidence drawn from what you actually said.   What makes it different The interview is built from your JD. Applying for an Android role and a data engineering role produce genuinely different interviews, because every question is generated against the extracted profile. The panel shares a memory. When the System Design interviewer takes over, it is handed what you told the Technical interviewer, plus the topics you already demonstrated, and told to build on them rather than re-test them. Routing is decided by code, not an LLM. A deterministic orchestrator tracks time spent, how deep it has gone on a topic, and answer quality. No model sits in that loop, so the interview cannot wander, cannot loop, and the routing adds no latency to a live conversation. It adapts to you. Answer a topic well and it goes deeper. Say you have never used something and it drops that topic and moves on. Each interviewer sounds like a different person. Per-persona voices are switched mid-session, so the handoff between rounds is something you hear, not something the transcript claims. How we use Agora Agora Conversational AI runs the conversation itself: listening, knowing when you have finished talking, letting you interrupt, and speaking back. Our FastAPI backend plugs in as its custom LLM, so every turn arrives as a request we control. That is where the choice of interviewer and the memory across rounds happen. Voice runs on Agora RTC, with AI echo cancellation added so interrupting still works on a phone speaker. Agora Signaling sends the live transcript, the interviewer's current state, and the timing of each speech recognition, model and voice step to the app.   Who it helpsCandidates, who can rehearse the interview for the role they actually want. And hiring teams: the same panel can screen candidates. Point it at your own job description and replace hours of engineer time on first-round phone screens with consistent evaluations backed by evidence.   Technical Architecture published in TowardsAI https://pub.towardsai.net/the-agent-boundary-problem-designing-multi-role-ai-systems-with-memory-and-orchestration-1ed09d58c345?source=friends_link&sk=e7771c32505a391676f3ad41060d5181