FinanceMan - Personal Finance Manager

A modern, responsive personal finance app built with Node.js, Express, Chart.js, Tailwind utilities, and vanilla JS. All data lives locally in JSON files for privacy and simplicity.

Features

• 📊 Insightful dashboard with charts, hero metrics, and page-specific panels
• ⚡ Global Quick Add button + Shift + A shortcut for instant transaction logging
• 🔥 Weekly Peek and daily streak chip to keep habits intact
• 🎉 Micro-celebrations (confetti) after every logged transaction
• 🎯 Goal tracker cards with editable targets
• 🔔 Browser notifications for streak reminders & upcoming bills
• 🏦 Multi-account management with inline balance editing
• 🏷️ Color-coded categories and smart budgeting controls
• 📁 Archive explorer with collapsible monthly history
• 🌓 Light/dark mode, fully responsive layout, CSV import/export, and more
• 🤖 Google Gemini-powered monthly summaries, Q&A, and smart category suggestions (all data stays on your machine)

Tech Stack

• Backend: Node.js + Express
• Frontend: HTML + Tailwind CSS + vanilla JS
• Charts: Chart.js
• Storage: Local JSON files (no external DBs)

Setup

git clone <repository-url>
cd financeman
npm install
npm start
# open http://localhost:3000

Note: FinanceMan stores everything in local JSON files. It’s designed to run on your machine (localhost) rather than a shared hosting environment.

For live reloading during development:

npm run dev

Gemini AI Setup

1. Create a free API key in Google AI Studio (Gemini free tier).

2. Copy the key into a .env file in the project root:

   GEMINI_API_KEY=your_key_here
   # Optional overrides:
   # GEMINI_MODEL=gemini-1.5-flash-latest
   # GEMINI_BASE_URL=https://generativelanguage.googleapis.com/v1

3. Install dependencies and run the server:

   npm install
   npm start
   # visit http://localhost:3000

4. From the dashboard, click "Generate AI Summary" under Gemini Monthly Coach or ask free-form questions in Ask Gemini About Your Money.

5. While adding a transaction, type a note and hit Suggest category to have Gemini choose the best matching category id.

6. If your key doesn’t have access to a specific model, leave GEMINI_MODEL unset; the server automatically falls back through current public flash variants (gemini-2.5-flash, gemini-2.5-flash-lite, gemini-2.0-flash, etc.).

Build Notes

• Crafted over a chill weekend (~6 focused hours) with endless lofi + coffee.
• Proudly built with help from OpenAI’s ChatGPT (GPT‑4 class) for ideation and speed boosts.
• Vibe code: "Track your stuff, celebrate the wins, keep it simple."

License

MIT