EchoCare

EchoCare is an AI-driven edge device voice-first assistive technology designed to empower blind and visually impaired individuals with real-time physical safety and complete digital accessibility. The system integrates obstacle detection, fall monitoring, and advanced GENAI-based voice assistance into one seamless solution that operates entirely hands-free.

EchoCare enhances physical safety through real-time obstacle monitoring and accelerometer-based fall detection, triggering an automatic SOS alert to whatsapp of their emergency contact if the user is unable to respond. Beyond physical safety, EchoCare enables blind users to independently access digital information through voice-enabled internet search, Gmail updates, Airbnb bookings, YouTube search with transcript summarization, finding locations and routes, and AI-powered legal document reading. Its built-in legitimacy checker identifies key points, hidden clauses, and suspicious content to protect users from fraud or exploitation.

The system is powered by a modern, scalable stack including Lovable, React, LivKit, Python FastAPI, MCP servers, Docker, Cloud Run, and advanced AI models such as Mistral VoxStral, ElevenLabs, Google Real-Time Model 2.0, and Mem0. Hardware integration uses accelerometer to detect fall and continuous situational awareness.

Designed for real-world reliability, EchoCare combines physical safety, digital empowerment, and AI reasoning to give blind individuals a level of independence that traditionally requires human assistance. It is practical, scalable, and immediately impactful making it a strong contender for real-world deployment and further innovation.

project links

https://github.com/jermiah/yosoup

https://github.com/KJFromMicromonic/ParisAIHackathon