Featured Project 1 stars

AI Portfolio Tracker & Financial News Hub

Real-time multi-asset financial dashboard integrated with Google Sheets, powered by pgvector semantic news search and Gemini AI insights.

Next.jsReact 19TypeScriptTailwind CSSSupabasepgvectorGeminiLangChainReact QueryAuth.js

View on GitHub Live Demo

Detailed Overview

A comprehensive, privacy-focused financial dashboard built with Next.js 16 to track, visualize, and analyze multi-asset investment portfolios (Equities, Bonds, Cash Flow). Integrates directly with personal Google Sheets as a headless database for real-time asset tracking. Features an automated financial news pipeline that ingests RSS feeds, translates and summarizes regional news using LangChain and Google Gemini, and performs high-speed semantic vector search via Supabase pgvector to automatically tag news relevant to your actual holdings.

Problem Statement

Generic portfolio trackers lock personal financial data into proprietary walled gardens, lack support for custom asset structures (such as bonds and daily transaction logs), and provide generic market headlines rather than contextualizing news to an investor's specific holdings.

Technical Solution

Engineered a personal financial hub that uses Google Sheets API v4 for complete data ownership and zero vendor lock-in. Integrated Google Gemini LLMs and Supabase pgvector to build an automated news engine that ingests RSS feeds, translates regional news to English, scores market sentiment, and dynamically links breaking financial stories directly to active portfolio holdings.

Architecture & Data Flow

Next.js 16 App Router frontend with React Query caching; Google Sheets API v4 for real-time portfolio data; automated RSS news ingestion pipeline powered by LangChain & Google Gemini for multilingual translation and sentiment extraction; Supabase PostgreSQL + pgvector storing embeddings with cosine similarity search (match_news); Auth.js (NextAuth v5) session security.