WealthSage AI

An autonomous AI financial architect with CRUD ledger control, royal themes, a glass mirror full-screen UI, and live quant analytics powered by Llama 3.3.

Comment

💡 Inspiration Managing personal finances is often overwhelming, and most budgeting apps are passive spreadsheets or dull dashboards that only show numbers without helping users fix their habits. We wanted to build more than just a tracker; we wanted an elite, data-driven financial architect. The inspiration for WealthSage AI Royal Edition was to create an intelligent, proactive agent with full CRUD database control that doesn't just read your data, but actively manages your ledger, automates subscriptions, and delivers brutal, actionable financial truths.

⚙️ What it does WealthSage AI is a full-stack, autonomous financial terminal wrapped in a stunning royal aesthetic with multiple luxury theme presets (Imperial Gold, Emerald Sovereign, Regal Burgundy, and Platinum Obsidian).

Key features include:

  • Glass Mirror Fullscreen AI Core: An immersive, translucent glassmorphic overlay for deep quant reasoning, natural language commands, and real-time ledger manipulation.
  • Autonomous CRUD Execution: Instead of just answering questions, WealthSage can execute real database actions—such as logging subscriptions, adding transactions, or resetting ledgers on command via Supabase.
  • Advanced Quantitative Analytics & Telemetry: Features an AI Executive Briefing, velocity scoring, compound portfolio engines, and tactical financial moves.
  • "What-If" Wealth Simulator: Interactive sliders projecting opportunity costs of daily expenses over time against S&P 500 compounding returns.
  • LaTeX Quant & Formula Notebook: Built-in research workspace supporting live Markdown and LaTeX mathematical proofs ($E=mc^2$).

🛠️ How we built it

  • Frontend: Next.js 14, React, TypeScript, and Tailwind CSS for a lightning-fast, highly responsive user interface with custom glassmorphism and dynamic theme variables.
  • Backend Services: FastAPI Python server orchestrating AI routing and background database synchronization.
  • AI & Intelligence: Powered by Llama 3.3 70B Versatile via the Groq API, using structured JSON output modes for reliable tool calling.
  • Persistence: Supabase PostgreSQL cloud database for live transactional security and storage.

🏗️ Challenges we ran into Balancing complex autonomous tool-calling loops with strict frontend rendering constraints was tricky. We solved this by structuring the Llama 3.3 model outputs into strict JSON schemas, allowing the backend to safely intercept database commands (like adding subscriptions or adjusting net worth) without breaking the UI stream.

🏆 What we've accomplished so far We successfully built a production-ready financial command center that merges high-end luxury aesthetics with rigorous quantitative finance tools, complete with instant demo persona loaders for judges.

📚 What we learned We learned how to build robust, agentic AI workflows that interact directly with relational databases safely, alongside mastering advanced Tailwind CSS variables for dynamic multi-theme customization.

🚀 What's next for WealthSage AI

  • Real institutional bank syncing via Plaid production webhooks.
  • Multi-currency support and automated tax-loss harvesting alerts.
  • Encrypted peer-to-peer wealth pooling and syndicate tracking.

Built With

Share this project:

Updates