SpendWise AI

AI Mom-style financial advisor that uses Plaid and Gmail to turn bank data into warm, actionable coaching. Stop overspending and build wealth with your personal AI Spending Personal.

Comment

About the SpendWise Project

The Genesis: A Universal Truth and a Market Gap

The inspiration behind SpendWise stemmed from a simple, yet profound, observation: everyone has a "Mom" figure in their life who, despite their best intentions, often knows more about their spending habits than they do themselves. This universal truth, coupled with the glaring market gap left by the discontinuation of popular personal finance tools like Mint, ignited our vision. We realized that while data-rich dashboards exist, they often lack the emotional intelligence, personalized accountability, and actionable guidance that truly changes financial behavior. We wanted to build an app that felt less like a cold spreadsheet and more like a trusted, albeit firm, advisor.

The Build: AI-Native, Secure, and User-Centric

Our approach to building SpendWise was rooted in leveraging cutting-edge AI and robust financial technology to create a truly unique user experience. The core of our system integrates:

Plaid API: For secure, read-only connections to over 12,000 financial institutions, ensuring real-time transaction data without compromising user credentials.

Anthropic Claude Sonnet: As our AI engine, Claude analyzes 90 days of transaction history to identify spending patterns, assign personalized "Spending Personas" (e.g., The Impulsive Foodie, The Subscription Hoarder), and provide nuanced financial advice.

Gmail API: To enable item-by-item analysis of Amazon and Costco receipts, offering unparalleled granularity in spending insights.

React 18, Vite, Node.js/Express: A modern, performant stack for a responsive and scalable application.

Every feature, from the "Ask Mom" conversational AI to multi-store price comparisons, was designed to be AI-native, ensuring that intelligence is not just an add-on but an intrinsic part of the product.

Challenges and Learnings: Navigating Complexity with Innovation

Developing SpendWise within a hackathon timeframe presented several significant challenges:

1. API Integration Complexity: Seamlessly integrating three powerful, yet distinct, APIs (Plaid, Anthropic, Gmail) required meticulous handling of authentication flows, data parsing, and error management. Ensuring secure, read-only access for Plaid was paramount.

2. AI Prompt Engineering: Crafting effective prompts for Claude AI to generate accurate spending personas, insightful transaction verdicts, and genuinely helpful "Mom-style" advice was an iterative process. Balancing warmth with honesty, and ensuring the advice was actionable, demanded careful refinement.

3. Real-time Data Processing: Analyzing 90 days of transaction data and performing real-time price comparisons required optimizing our backend logic for efficiency and speed, especially given the diverse data structures from various sources.

4. User Experience Design: Translating the "Mom" persona into an intuitive and engaging user interface was crucial. We aimed for a design that felt personal and trustworthy, avoiding the sterile feel of many traditional finance apps.

Through these challenges, we learned the immense power of focused integration and iterative AI development. We discovered that the most impactful solutions often lie at the intersection of advanced technology and deep human understanding. The hackathon pushed us to rapidly prototype, validate, and refine our core concept, proving that a truly intelligent and empathetic financial advisor is not only possible but deeply needed.

The Vision Forward

SpendWise is more than just an app; it's a movement towards financial literacy and empowerment, delivered with a touch of familiar wisdom. We believe that by providing users with clarity, accountability, and personalized advice, we can help millions transform their relationship with money, one "Mom-approved" decision at a time.

Built With

Share this project:

Updates