How I Revived VoiceAssist AI: Adding Kannada & Urdu Support with GitHub Copilot

A developer used GitHub Copilot to expand VoiceAssist AI, a multilingual voice assistant for Indian bank counters, from 6 to 8 languages by adding Kannada and Urdu support. Copilot automatically generated a pull request that implemented Devanagari and Persian/Arabic script detection, TTS fallback handling, and updated the frontend language list. The update enables voice input for both languages via Whisper STT and includes a full verification flow.

VoiceAssist AI is a multilingual voice assistant I built for bank branch counters during the iDEA 2.0 hackathon. It lets customers speak in their native language while bank staff respond in English — bridging the language gap in Indian banking. The app supported 6 languages but was missing Kannada and Urdu — two major Indian languages covering millions of bank customers. The code also had no proper script detection for non-Latin scripts. I asked Copilot to analyze the repo and suggest what was missing. It identified the language gaps, then automatically created a PR adding Kannada Devanagari script detection, Urdu Persian/Arabic script detection, TTS fallback handling, and updated the frontend language list from 6 to 8 languages. Languages expanded: 6 → 8 Kannada + Urdu voice input via Whisper STT Proper Unicode script detection Full verification flow in both languages GitHub: https://github.com/rohan-jadhav-dev/voiceassist-ai https://github.com/rohan-jadhav-dev/voiceassist-ai Live: https://voiceassist-ai.vercel.app https://voiceassist-ai.vercel.app