Wexa AI

• 

  logo

Inspiration Financial management tools are historically broken: they are either bogged down by complex corporate jargon, cluttered with confusing data tables, or hidden behind high-friction paywalls. Most users give up on personal finance apps simply because the onboarding curve is too steep.

Wexa AI was created by Yash Choubey to eliminate this friction entirely., Wexa AI operates under a foundational principle: The 5-Year-Old Rule. The core vision is to translate high-level financial metrics, stock market fluctuations, and security audits into instant visual signals—such as intuitive Traffic Light Status Badges (🟢 Operational, 🟡 Warning, 🔴 Critical)—allowing any user to assess their financial health in seconds without needing an accounting degree.

What it does Wexa AI functions as an end-to-end, autonomous financial intelligence platform structured into four primary modules:

📸 MAGIC TOOLS (Gemini Vision Receipt Scanner): Users can upload receipts, invoices, or utility bills. The system instantly parses the image, extracts merchant names, calculates totals, categorizes expenses, and flags hidden fees or suspicious charges.

🌤️ MARKET WEATHER (Real-Time Stock Ticker & Analytics): Displays live global market trends, stock marquees, and interactive charts with a visual "weather" theme (e.g., sunny for bullish, stormy for market dips).

🐷 MY PIGGY BANK (AI Wealth Planner & Budgeting): An interactive personal financial assistant that ingests user income and goals to provide real-time saving recommendations, budget tracking, and custom financial advice.

🛡️ COMPANY & AUDITS (Security & Export Suite): Monitors system performance, security audit logs, and eco-efficiency scores, allowing users to export full report summaries in CSV or printable PDF formats with a single click.

How we built it Wexa AI combines modern 3D web technologies, real-time market APIs, and Google’s latest multimodal AI models:

Frontend Architecture: Built using React, Vite, and Tailwind CSS to achieve a modern, dark-mode startup aesthetic with sub-second page transitions.

Interactive 3D Graphics: Powered by Three.js, @react-three/fiber, and @react-three/drei to render interactive, hardware-accelerated financial charts and 3D visual canvas elements.

AI Core Integration: Leverages Google Gemini 3.1 Pro for natural language financial analysis and Gemini Vision OCR for receipt auditing and text extraction.

Market Data Pipeline: Connected to the Twelve Data API to pull real-time stock ticker values and historical market candlestick charts.

Payment Engine: Integrated with the Instamojo API to handle Pro level subscription upgrades ($9/mo) via secure signature validation.

Deployment & CI/CD: Live deployment hosted on Vercel (wexa-ai-cwy.vercel.app) with continuous automated builds.

Challenges we ran into API Rate-Limiting & High Traffic: Continuous streaming of stock market data caused frequent HTTP 429 rate-limit errors from third-party APIs during peak usage. To solve this, a custom 60-second local memory caching layer was implemented alongside smooth fallback mock states, ensuring the UI remains continuous and functional.

Monetization Refactoring: Transitioning away from legacy payment hooks to a native Instamojo integration required building secure backend payload signing without interrupting the client-side user experience or breaking application state.

OCR Precision on Unstructured Receipts: Extracting accurate total amounts from noisy, low-light, or crumpled receipt photos required fine-tuning prompts for Gemini Vision OCR and building robust fallback parser logic to prevent app crashes.

Accomplishments that we're proud of 95%+ Production Readiness: Maintained zero TypeScript or bundling errors on live Vercel deployments, achieving a production-grade build ready for real-world traffic.

High-Performance 3D Visuals: Successfully integrated Three.js graphics that maintain smooth 60 FPS performance on both mobile devices and desktop screens.

Radical UI Simplification: Fully replaced complex financial terminology with intuitive visual cues (Traffic Light badges and dynamic icons) without losing numerical accuracy.

What we learned Designing for Simplicity is Hard: Engineering a UI around "The 5-Year-Old Rule" requires significantly more discipline than building traditional data-dense dashboards. Every visual element must serve an immediate purpose.

Resilient Client Architecture: Combining multiple external data streams (Gemini AI + Twelve Data + Instamojo) requires strong client-side caching, error boundaries, and defensive state management to prevent cascading API failures.

What's next for Wexa AI Multi-Currency Vision OCR: Extending Gemini Vision capabilities to automatically detect and convert foreign currencies on international receipts.

Collaborative Family Piggy Banks: Introducing shared multi-user budgeting drawers for families and small business teams.

Automated Banking Webhooks: Partnering with open-banking APIs for direct, real-time sync with user bank accounts.

Built With

• gemini-vision
• google-gemini-api
• instamojo-api
• react
• react-three-fiber
• tailwindcss
• three.js
• twelve-data-api
• typescript
• vercel
• vite