Inspiration

The idea for Yojna Mitra was inspired by a simple observation: many government welfare schemes (more than 5,000 in India) exist to support citizens, yet people often struggle to discover and understand them. We realized that information barriers, language barriers, and complex application processes can prevent eligible citizens from accessing benefits that could improve their lives. We wanted to explore how conversational AI could make welfare information easier to understand, more accessible, and more personalized.

What it does

Yojna Mitra is a multilingual, voice-based AI assistant that helps underserved Indian citizens discover government welfare schemes through simple conversations. A rural farmer seeking agricultural support, a student searching for scholarships, or an elderly citizen applying for a pension can simply describe their needs in their preferred language.

Although hundreds of government welfare schemes exist across education, healthcare, agriculture, housing, employment, women empowerment, and social protection, many eligible citizens never benefit because information is difficult to find and written in complex language. These challenges are especially significant for citizens with limited literacy, limited digital skills, or limited English proficiency.

Yojna Mitra addresses this gap through a voice-first, conversational experience. Users can speak or type their needs, select their preferred language, and answer a few simple eligibility questions related to factors such as age, income, occupation, education level, and location. The AI then recommends relevant schemes, explains benefits and eligibility criteria in simple language, provides application guidance, and directs users to verified official resources from myScheme.gov.in.

Unlike a traditional search portal, Yojna Mitra allows users to describe their situation in natural language. The AI understands user intent, asks personalized follow-up questions, and simplifies complex eligibility information into actionable guidance.

How we built it

We built the solution using React, TypeScript, Vite, Capacitor, Android, Node.js, Google Gemini, Google Cloud Run, Google Secret Manager, Speech Recognition, and Text-to-Speech technologies. The application supports multilingual conversations, voice input, spoken responses, category-based browsing, and personalized scheme recommendations. The backend securely connects to Gemini through Cloud Run and Secret Manager, ensuring API credentials are never exposed in the mobile application.

Challenges we ran into

One of our biggest challenges was building a reliable multilingual voice experience while ensuring secure AI access and preventing inaccurate scheme recommendations. We also needed to make speech recognition and text-to-speech work smoothly across different languages, design an intuitive interface suitable for first-time smartphone users, and ensure that recommendations were based on verified government information rather than potentially hallucinated AI responses.

Accomplishments that we're proud of

We are proud of developing a working AI-powered solution that combines accessibility, personalization, and verified government information to address a real social problem. We successfully built a voice-first experience that supports regional languages, provides personalized welfare scheme recommendations, and makes complex government information easier to understand for underserved communities.

What we learned

Through this project, we learned the importance of responsible AI design, user-centered experiences, and combining AI capabilities with trusted public data sources. We also gained experience in building secure AI-powered applications, designing conversational user interfaces, integrating speech technologies, and ensuring that AI systems remain transparent and helpful rather than acting as decision-makers.

What's next for Yojna Mitra: AI-Powered Welfare Scheme Discovery Assistant

In the future, we plan to expand the platform beyond scheme discovery by helping users not only identify suitable welfare schemes, but also complete application processes, organize and upload required documents, and receive personalized assistance throughout their journey. We also plan to improve regional language coverage, strengthen recommendation accuracy, and support personalized user profiles.

Have uploaded APK under Additional Info section

Built With

Share this project:

Updates