Building Bhasha Academy: A Multi-Agent Hinglish Voice Tutor with Murf Falcon & LiveKit
Building Bhasha Academy: An AI Voice Tutor for English and Math Language is meant to be spoken, not just read. For millions of students and job seekers in India, being confident in English and basic math can open the door to better career opportunities. But many learners don't have access to a personal tutor. Even when they do, the fear of making mistakes in front of someone can make practicing…
Bhasha Academy is an AI-powered voice tutor designed to provide learners with a private, judgment-free space to practice speaking English and solve basic math problems. The system is built using several services, with LiveKit at the center of the real-time voice experience. Users interact with the system through a voice-first interface, allowing them to simply say, "Mujhe English practice karni hai" or "Samar, samay ka 25 percent kitna hai?" to have the agent understand their request and respond naturally.
The system supports Hinglish, an English-Hindi mixed language, and uses Indian-accented voices for a more natural experience. The main components of the architecture include LiveKit, speech-to-text using Deepgram Nova-3, an LLM (Google Gemini 3.5 Flash Lite) for conversation understanding and generation, Murf Falcon TTS for voice generation, and SQLite for storing learner profiles and learning progress.
The system employs specialist agents, Anisha for general language practice and Samar for math practice, enabling seamless handoffs between agents to maintain context. Additionally, Bhasha Academy supports phone/SIP calls and voicemail detection, enabling scheduled lessons and human teacher escalation when needed.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
