Polara

Inspiration

A stark truth inspired our project: while tens of billions of dollars in government benefits are available to retired, disabled, and veteran citizens, millions miss out because the systems are too complex or confusing. For instance, nearly 60% of eligible low-income older adults do not claim SNAP food benefits. We were determined to build a solution that replaces bureaucratic hurdles with a tool defined by clarity, accessibility, and speed. We chose the Solar System metaphor to provide users with a clear, holistic view of their entire financial and benefit cosmos.

What it does

Polara is a voice-driven, multilingual platform that manages and optimizes benefits by visualizing a user's entire financial situation using an accessible space metaphor. Each planet represents a major benefit category (Healthcare, Housing, Food, etc) and the Sun is the central financial hub for SSI/SSDI and Taxes. Polara's core function is to provide a single, interactive dashboard that shows the value and status of all benefit categories, allowing users to navigate and act on financial opportunities simply by speaking commands to the integrated AI assistant.

How we built it

We implemented a full-stack architecture prioritizing modern APIs and accessibility standards, using React (Vite) with TailwindCSS and Framer Motion for a sleek, accessible frontend. For the core innovation, we leveraged the Eleven Labs API for high-quality voice-to-text narration and the Gemini API for robust, multilingual AI intent classification to execute voice commands. The backend runs on Node.js with Express, using a flexible MongoDB database for the data model, secured by JWT authentication and AES-256 encryption for sensitive financial data, while react-i18next handles dynamic content updates across our 7 supported languages.

Challenges we ran into

The greatest technical challenge we faced was ensuring voice capture, narration, and unique accessibility features (like color-blindness filters) worked flawlessly across diverse devices and configurations. Integrating high-fidelity audio from the Eleven Labs API for narration and microphone input demanded rigorous cross-browser and device testing, as coordinating the audio stream with the dynamic UI, especially when High Contrast Mode or the Dyslexia Font was active, often led to unexpected synchronization issues that required constant tuning.

Accomplishments that we're proud of

Our greatest accomplishment is delivering on our core promise of radical accessibility for an underserved population. We are most proud of successfully transforming five complex, interlocking benefit categories into the intuitive, single-screen Solar System visual map, which immediately clarifies the user's financial picture. Finally, successfully integrating the Gemini API allowed us to create a truly functional voice assistant capable of handling and classifying policy-heavy commands across multiple languages, dramatically lowering the barrier to access.

What we learned

We learned that achieving an effortless user experience requires a deep understanding of human-centered engineering. Our primary lesson was the complexity of the voice stack. We learned how to precisely integrate the Eleven Labs API for high-quality text-to-speech and the Gemini API for voice command classification, ensuring seamless, multilingual audio recognition despite background noise and varied accents. This challenge forced us to manage frontend state and backend logic to abstract away dense government rule sets. We also learned the importance of intuitive design and inclusive color theory firsthand by implementing features like the color-blindness filters, high-contrast modes, and the dyslexia font, ensuring that the Solar System UI was visually accessible, legible, and functional for all users, not just those with perfect vision.

What's next for Polara

Our next steps are focused on leveraging data intelligence to get ahead of user needs. Our primary goal is the rollout of predictive benefit modeling, which will transform Polara from a reactive management tool into a proactive financial assistant. We will first develop connectivity to IRS and SSA sandbox APIs to allow users to securely and instantly pre-fill complex financial data, minimizing manual entry. This data will then fuel our machine learning models, enabling Polara to proactively alert users that they are about to qualify for a benefit based on projected changes in income, age, or health status. Finally, we will finalize the architecture for a dedicated mobile application, ensuring our design is optimized for users on fixed incomes who rely heavily on mobile devices.