OpenShark
Financial anxiety is simply the fog of not being able to see your true numbers.
Core trick
Pulls bank transactions through Plaid into SQLite, runs anomaly and recurring-subscription detection, then has an LLM write insights.
Description
Financial anxiety is simply the fog of not being able to see your true numbers. Manual tracking is a low-leverage trap that keeps you busy but poor. OpenShark fixes that
README excerpt
# 🦈 OpenShark: AI-Powered Fintech Intelligence OpenShark is a next-generation fintech platform built for the Google Hackathon. It combines real-time banking data (via Plaid) with an advanced AI insights engine to provide users with a "Financial DNA" analysis, anomaly detection, subscription tracking, and cashflow forecasting.  ## 🏗 Project Architecture The project is divided into three main components: - **Backend (`/backend`)**: A FastAPI (Python) server that manages the SQLite database, handles Plaid integrations, and runs the analytical insights engine. - **Mobile App (`/my-fintech-app`)**: A cross-platform React Native (Expo) application providing a sleek, Material Design-inspired dashboard for transactions and AI insights. - **AI Agent (`/agent-node`)**: Powered by **Ledger**, an autonomous financial forensic analyst. Ledger uses `pi-agent-core` to analyze raw data, detect "leaks" in cash flow, and provide dry, professional, and action-oriented financial advice directly from the database. ## ✨ Key Features - **🏦 Plaid Integration**: Seamlessly connect bank accounts to sync real-time transaction data. - **🧠 AI Insights Engine**: - **Financial DNA**: Categorizes spending patterns. - **Anomaly Detection**: Flags unusual transactions. - **Subscription Tracking**: Identifies recurring charges (Netflix, Spotify, etc.). - **Cashflow Forecasting**: Predicts future balances based on historical trends. - **🎙 Narrated Insigh