🌾 AgriAlert (कृषीअलर्ट) — Voice AI for Indian Farmers

AgriAlert is a voice AI assistant built for Maharashtra's farmers. It delivers crop advisory, real-time weather alerts, mandi prices, and agricultural guidance — all through natural Marathi conversation. Built for the #VoiceForBharat Farm & Field track, powered by Murf Falcon TTS and LiveKit.

---

✨ Features

• 🗣️ Marathi Voice Conversations — Natural Devanagari speech via Murf Falcon (Pooja voice)
• 🌤️ Live Weather Alerts — Real-time forecasts from Open-Meteo based on the farmer's district
• 💰 Mandi Price Lookup — Crop market prices by district (mock data, extensible)
• 🧠 Caller Memory — SQLite-backed memory remembers returning farmers (name, crop, district)
• 🔀 Specialist Handoff — Seamlessly transfers complex crop disease queries to an expert agent
• 🚨 Human Escalation — Creates support tickets for unresolved issues with consent
• 📲 Real-time UI — Live data cards pushed to the frontend via LiveKit DataChannels
• 🔒 Privacy-first — Explicit consent required before saving any personal data

---

Architecture

flowchart LR
    A[🎙️ User speaks] -->|audio| B[Deepgram STT]
    B -->|text| C[Gemini LLM]
    C -->|response text| D[Murf Falcon TTS]
    D -->|audio| E[LiveKit]
    E -->|stream| F[🔊 User hears]

    style A fill:#444441,stroke:#888780,color:#fff
    style B fill:#185FA5,stroke:#85B7EB,color:#fff
    style C fill:#534AB7,stroke:#AFA9EC,color:#fff
    style D fill:#0F6E56,stroke:#5DCAA5,color:#fff
    style E fill:#D85A30,stroke:#F0997B,color:#fff
    style F fill:#444441,stroke:#888780,color:#fff

Loading

---

Why Murf Falcon

• 55ms model latency — fastest production TTS
• 130ms time-to-first-audio across 10+ global regions
• $0.01/1000 characters — up to 10x cheaper than alternatives
• 150+ voices across 35+ languages
• 99.38% pronunciation accuracy

---

Quickstart

Prerequisites

• Python 3.10+
• uv — fast Python package manager

  # macOS/Linux
  curl -LsSf https://astral.sh/uv/install.sh | sh
  # Windows (PowerShell)
  powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

• Node.js 18+
• pnpm — fast Node package manager

  npm install -g pnpm

• A LiveKit project (free tier available)

Step 1: Clone the repo

git clone https://github.com/code-with-parth/AgriAlert.git
cd AgriAlert

Step 2: Set up environment variables

Create .env.local in both backend/ and frontend/ (copy from .env.example in each). You need:

Variable	Where to get it	Required
LIVEKIT_URL	LiveKit Cloud dashboard	Yes
LIVEKIT_API_KEY	LiveKit Cloud dashboard	Yes
LIVEKIT_API_SECRET	LiveKit Cloud dashboard	Yes
MURF_API_KEY	murf.ai/api/dashboard	Yes
DEEPGRAM_API_KEY	deepgram.com	Yes
GOOGLE_API_KEY	Google AI Studio	Yes

Step 3: Install & run

Option A — All-in-one (from repo root):

# macOS/Linux
chmod +x start_app.sh && ./start_app.sh

# Windows (PowerShell)
.\start_app.ps1

Option B — Separate terminals:

# Terminal 1 — Backend agent
cd backend && uv sync && uv run python src/agent.py download-files
uv run python src/agent.py dev

# Terminal 2 — Frontend
cd frontend && pnpm install && pnpm dev

Open http://localhost:3000, click संवाद सुरू करा / Start Conversation, allow microphone access, and speak.

---

Project Structure

AgriAlert/
├── backend/                 # Python voice agent (LiveKit Agents + Murf Falcon)
│   ├── src/
│   │   ├── agent.py         # Main agent — pipeline, system prompt, tools
│   │   ├── crop_specialist.py # Specialist agent for deep crop queries
│   │   └── db.py            # SQLite caller memory & analytics
│   ├── tests/               # LLM-judged evaluation tests
│   ├── .env.example         # Backend env template
│   └── pyproject.toml       # Python deps (uv)
├── frontend/                # Next.js UI for voice sessions
│   ├── app/
│   │   ├── page.tsx         # Main voice page
│   │   ├── dashboard/       # Analytics dashboard
│   │   └── api/token/       # LiveKit token endpoint
│   ├── components/          # UI (agents-ui, app config, theme)
│   ├── app-config.ts        # Branding, accent colors, visualizer
│   └── package.json         # Node deps (pnpm)
├── start_app.sh             # Start all services (macOS/Linux)
├── start_app.ps1            # Start all services (Windows)
└── README.md

---

Configuration

Voice

Edit the tts=murf.TTS(...) call in backend/src/agent.py:

tts=murf.TTS(model="falcon", voice="Pooja", locale="mr-IN", style="Conversation")

Browse all voices: Murf Voice Library

LLM

Default is Gemini (gemini-3.5-flash-lite). To switch to OpenAI, set OPENAI_API_KEY and update the llm= call in agent.py.

STT

Default is Deepgram Nova-3 with Marathi (language="mr"). Configurable in the AgentSession(stt=...) call.

---

Deploy

Backend → Railway

Frontend → Vercel

Both services connect via LiveKit — use the same LIVEKIT_URL, LIVEKIT_API_KEY, and LIVEKIT_API_SECRET on both platforms.

---

Links

• Murf API Docs
• Murf Voice Library
• LiveKit Agents Docs
• Deepgram Docs

---

License

MIT