

📊 Day 8 of 10 Days of Voice Agents — Voice for Bharat Edition 🇮🇳🌾
I am building a voice agent using the fastest TTS API — Murf Falcon.
For Day 8 (Farm & Field Track), I built a real-time "Call Analytics & Performance Dashboard" for Kisan Vaani — connected directly to SQLite (kisan_memory.db) to track every voice call outcome, success rate, regional language, and duration.
🎙️ What I built on Day 8:
• 🌾 Defined Success Criteria for Kisan Vaani: A call is marked SUCCESSFUL when requested mandi rates/weather information is delivered, saved SQLite memory is loaded, or an emergency KVK officer ticket is created. Marked FAILED if disconnected prematurely before query completion.
• 🗄️ SQLite Session Logger (call_logs table): Automatically logs every call session upon disconnect with Call ID, Caller Name, District, Language, Outcome Summary, Duration (sec), and Status (SUCCESS / FAILED).
• 🌐 Multilingual Regional Support: Added explicit language tracking (Hindi Devanagari, Kannada, Marathi, Punjabi, and Tamil) to monitor regional voice accessibility across Bharat!
• 📊 Real-Time Web Analytics Dashboard UI: An interactive glassmorphism dashboard featuring 4 live metric cards:
📈 Total Calls
✅ Successful Calls
❌ Failed Calls
🎯 Success Rate %
• 📜 Protected Call History Log Table: Renders recent calls with colour-coded status badges, regional language badges, time, district, and outcome summaries while protecting caller privacy.
• 🔄 Live Increment Verification: Completing a successful call automatically increments Total Calls and Successful Calls live on the dashboard!
Watch the attached demo video to see a live call incrementing metrics on the analytics dashboard 🎥
💻 The GitHub repo link is in the FIRST COMMENT. 👇
Thank you, Murf AI & LiveKit, for organising the #VoiceForBharat challenge.
#VoiceForBharat #10DaysofAIVoiceAgents #MurfFalcon #MurfAI #LiveKit #AgriTech #AIagents #BuildInPublic #VoiceAI #CallAnalytics #SQLite #GenerativeAI
13 August 2026 à 18h23
View on LinkedIn →