Building Suraksha Sathi: My 10-Day Journey into Voice AI for Disaster Response
🌪️ Building Suraksha Sathi: A Hindi/Hinglish Voice Agent for Responsible Disaster Response Building a Hindi/Hinglish voice agent designed to provide responsible disaster-response assistance through real-time voice conversations. What if, during a disaster, someone could simply speak to an AI assistant instead of navigating complicated interfaces? That question became the starting point for my…
Title: Creating Suraksha Sathi: A 10-Day Journey into Hindi/Hinglish Voice AI for Disaster Response
During the 10 Days of Voice Agents challenge, I embarked on an ambitious project to develop a Hindi/Hinglish voice agent called Suraksha Sathi. This innovative system aimed to provide responsible disaster-response assistance through real-time voice conversations. The concept was simple: during a disaster, someone could simply speak to an AI assistant, rather than having to navigate complex interfaces.
The project evolved over ten days, starting from a basic voice conversation and gradually transforming into a comprehensive disaster-response voice system. Key features included real-time voice conversations, support for Hindi, English, and Hinglish, persistent user memory, disaster-response tools, outbound welfare calls, human escalation, call analytics, specialist-agent handoffs, and safety-focused disaster-response guardrails.
The inspiration behind building Suraksha Sathi was to create a conversational layer for disaster-response information and assistance. In emergencies, quick access to accurate information can be vital. Traditional methods of seeking information, such as typing through menus or searching multiple websites, may not always be convenient, especially during high-stress situations. Voice provides an alternative interface that allows users to communicate naturally.
The project's architecture gradually evolved into a multi-component voice-agent system. The main components included the user interface, real-time communication, Suraksha Sathi (the main agent), memory, tools, human escalation, call analytics, specialist agents (such as ShelterSathi for shelter information), Murf Falcon (text-to-speech), and a SIP trunk for outbound calls.
One of the most important aspects of Suraksha Sathi was its commitment to safety. Given the critical nature of disaster response, the agent had to be extra cautious about what it claimed. It was programmed to never issue evacuation orders, declare areas safe, claim that a disaster had ended, invent emergency alerts, shelter availability, or capacity, or assert that someone had been rescued or that emergency responders had been dispatched unless verified.
Instead, the agent focused on providing verified information and general safety guidance, distinguishing between what it knew to be true and what it couldn't confirm.
Key lessons learned during the challenge emphasized the importance of knowing what not to say in safety-critical systems, just as much as knowing what to say. The project aimed to strike a balance between providing helpful information and maintaining a cautious approach to avoid spreading misinformation during emergencies. Suraksha Sathi was designed to be a responsible conversational tool, guiding users towards the most relevant and accurate disaster-response assistance available.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
