Finzytrack

Open-source personal finance software for people who want to understand their finances without giving up privacy or control. Built on Beancount double-entry bookkeeping — your data lives in a plain-text ledger file on your machine, readable by any text editor, with no lock-in. AI is entirely optional and works with any model you choose.

Features

• Double-entry bookkeeping — built on Beancount, the powerful plain-text accounting system
• Import a wide variety of statements — rule-based importers for OFX, CSV, XLS files, PDF statements, or directly from email
• Find any transaction, instantly — global search across your full history, with advanced filtering by date, amount, account, and more
• Customizable dashboards — KPI cards, charts, tables, and pivot tables in configurable grid layouts
• Auto-categorization — train on your past transactions or use AI assistance
• Powerful querying — SQL and BQL queries against your financial data
• Optional AI assistance — parse statements, create import rules, enter transactions in natural language, build dashboards, and answer financial questions
• You own your data — everything stored in a single plain-text Beancount ledger file with zero lock-in

Documentation

Full documentation at docs.finzytrack.com, including:

• Quick Start — get up and running from first launch
• Installation — download and install on macOS (Apple Silicon), Linux, or Windows
• Views — guide to each screen in the app
• Reference — technical reference for recipes, rules, queries, and configuration

Development Setup

See Building from Source for full development instructions. But here's a quickstart — first, install the platform-specific prerequisites for your platform (on Linux, without them pip install will fail building PyGObject). Then:

git clone git@github.com:sagarbehere/finzytrack.git
cd finzytrack

# Backend
python3 -m venv venv
source venv/bin/activate
cd backend
pip install -r requirements.txt

# Frontend
cd ../frontend
npm install

Contributing

Bug reports and feature requests are very welcome — please open a GitHub issue and I'll do my best to address them.

I'm not currently accepting pull requests. This is a personal side project, and I can't commit to the kind of timely, thoughtful code review that contributors deserve. I'd rather be upfront about that than leave patches sitting unreviewed for weeks. If you've found a bug or have an idea, please open an issue and I'll pick it up when I can.

License

This project is licensed under the GNU General Public License v2.0.