Project Karunē

A technology-driven platform that serves as a bridge and connection between what people need with what people can give, at scale.

Comment

Inspiration

Food waste, loneliness, and lack of access are not isolated problems they are symptoms of disconnection. Through on-ground experience working with shelter homes, we saw that help exists, but it often doesn’t reach the right place at the right time. Millions of children grow up without mentorship or guidance, and elderly individuals feel forgotten not because people don’t care, but because systems don’t connect them efficiently. Karunē Connect was inspired by the idea that compassion should scale the way technology does.

What it does

Karunē Connect is a live web platform that connects shelter homes, volunteers, and donors in one transparent ecosystem. Shelter homes can post real-time needs such as meals, groceries, mentorship sessions, or skill-building programs while volunteers and sponsors can discover and respond to these needs based on relevance and proximity. The platform also includes an AI-powered call agent that allows shelter homes to place requests via phone in regional languages of India as well as English, ensuring accessibility even without internet access.

How we built it

Karunē Connect was built using a modular, scalable tech stack. The frontend is developed using HTML and JavaScript, with Firebase handling authentication and backend services. EmailJS is used for automated email communication, Google Maps API enables location-based matching, and the platform is hosted using GitHub Pages with a GoDaddy domain. We implemented a rule-based prioritization algorithm that ranks requests based on proximity ensuring that volunteers and sponsors closer to a shelter see those needs first. In parallel, we developed an AI recommendation model in Python that factors in urgency, skills, and historical data to generate explainable recommendations. While the model is functional and running locally, cloud deployment and API integration are planned as the next step. To improve accessibility, we built an AI-powered call agent using Vapi.

Challenges we ran into

One of the biggest challenges was integrating an AI model into a live platform within the constraints of a two-day hackathon. Deploying AI models as APIs requires cloud infrastructure, orchestration, and investment beyond the scope of the event. Another challenge was designing a system that is inclusive for users who may not be digitally fluent or English-speaking, which led to the development of the AI call agent.

What we learned

We learned how to balance technical ambition with real-world constraints, build scalable MVPs, and design for inclusivity. As a team, we collaborated across domains, combining social problem understanding with frontend development, backend logic, AI modeling, and conversational AI systems.

What's next for Project Karunē

Post-hackathon, we plan to formally register Project Karunē as a nonprofit organization to ensure structured, accountable impact. On the technical side, we will deploy the AI recommendation model using cloud infrastructure, expand the AI call agent to more regional languages, and scale the platform city-by-city starting with Delhi, Hydrabad and Mumbai before expanding globally. Our goal is to make compassion systematic, scalable, and accessible.

Built With

Share this project:

Updates