• 

  FinDigi Landing page

• 

  Mock Data Sample

The Story Behind FinDigi

Curiosity has always been the fuel behind my best ideas. FinDigi was born from one such spark.

Inspiration

While working at a fintech company in Nigeria, I constantly witnessed the daily struggles users faced in understanding and managing their finances. We had access to incredible amounts of data—transactions, behaviors, spending trends—but making sense of it all was an entirely different ball game.

One day, a heated internal debate arose: “Should we integrate AI to provide users with actionable financial insights?” The idea was exciting, but the question was: Which AI service? How do we personalize finance at scale, and with what engine?

Then I stumbled upon the [Perplexity Sonar Hackathon]. That was my "aha!" moment. I saw it as an opportunity to prototype the very vision we had debated—an intelligent, user-centered digital finance assistant. And so, FinDigi was born.

What it does

FinDigi is an all-in-one personal finance platform. It doesn't just track your spending — it learns from it, simulates your future, advises you in real-time, and explains the why behind every financial fluctuation. It is a smart, interactive, and privacy-first personal finance web application designed to help users analyze, simulate, and visualize their financial journey — without a bank account. Built with AI using the Sonar API, FinDigi empowers users with data-backed insights, mock investing, and personalized financial forecasting.

How I built it

FinDigi has four major modules, each powered by a different model from the Sonar API:

1. BudgetBuddy (powered by sonar-pro)

Allows users to upload bank statements and get categorized insights, spending patterns, and savings advice in real time.

2. iAdvisor (powered by sonar-deep-research)

Delivers high-level investment reports and strategies based on user goals—like a personal AI wealth advisor in your pocket.

3. InvestiLoop (powered by sonar-reasoning)

An investing simulator where users get live AI feedback on pretend investments. Think of it as a safe playground to learn finance with AI as your coach.

4. FinTwin (powered by sonar-reasoning)

Builds a "financial digital twin" of the user to simulate their 1, 5, and 10-year financial future—and even compares their trajectory to their peers.

All insights were generated via endpoints that called specific Sonar models, while the frontend (built with React and styled with Tailwind CSS) displayed rich visualizations. State and sessionStorage were used to preserve user data across sessions.

Challenges I ran into

Thankfully, I didn't face any major roadblocks. The only minor hiccup was handling the Sonar API response format, especially for insights wrapped in markdown. I solved this by using the marked package to clean and parse the results into safe HTML content.

What I learned

Exploring the Sonar API opened my eyes to the true capabilities of personalized AI. Not only does it provide contextual, tailored insights, it also allows developers to choose specific models for different levels of reasoning and depth.

This project taught me:

• How to prompt engineer for high-quality, nuanced financial insights.
• The importance of modular AI usage (matching the right Sonar model to the right use-case).
• How a clean UX backed by intelligent responses transforms how users experience financial literacy.

What's next for FinDigi

I plan to evolve FinDigi into a production-ready fintech solution. Here's what’s next on the roadmap:

• Tighten Security
  Implement proper backend authentication with JWT, OAuth, and 2FA for secure user sessions.

• Handle Diverse CSVs
  Expand support for various CSV formats from banks and fintechs, including messy or mixed debit/credit logs.

• Mobile App Version
  Build native iOS/Android versions with biometric authentication and offline access.

• Deeper Data Analysis
  Go beyond basic stats — uncover spending anomalies, recommend budgeting changes, and track financial goals.

• Hashed Requests/Responses
  Secure all client-server communication using encrypted payloads and request validation mechanisms.

• Fintech Pairing Support
  Allow users to sync FinDigi with their fintech apps or banks to automate data import.

• Conversational AI Assistant
  Add an AI chatbot to explain insights, answer finance questions, and guide users through their financial reports.

• Localization & Currency Support
  Support for multi-currency conversions and localized finance insights across regions.

• Goal-Based Planning
  Let users set savings, emergency funds, or investment goals and track progress with intelligent reminders.

• PDF & Report Export
  Generate printable/downloadable summaries and analysis for finance tracking or compliance purposes.

• Collaboration Tools
  Enable users to share analysis with financial advisors, partners, or family members securely.

Final Thoughts

FinDigi was not just a project; it was an agreement to what’s possible when AI meets real user needs. What started as a curiosity-driven experiment ended up as a modular, scalable personal finance assistant—built on the backbone of an API that truly understands context.

I'm excited about the future possibilities: integrations with real banks, more behavioral modeling, and even ethical financial nudging powered by AI.

With tools like Sonar, the future of finance is no longer cold numbers—it’s intelligent, humanized, and personal.

Built With

• dayjs
• framer-motion
• lucide-react
• marked
• mui
• papaparse
• react
• react-chartjs
• sonner
• tailwind
• uuid