99%
Document parse accuracy
Global
Tax coverage
+40%
Goal completion
270
Canvas frames at 60fps
Most finance tools tell you what you spent. FinOS tells you what you can do next. I built a simulator that ingests real documents, understands them, and projects a plan around the numbers.
Documents enter through an OCR pipeline. A tax engine localises deductions to your country and region. A roadmap generator converts net surplus into sequential, compounding goals, each with its own AI-assisted workspace.
- Next.js
- React
- Framer Motion
- MongoDB
- Gemini
- Tesseract.js
Reads your financial documents, models tax rules for countries around the world, and turns the surplus into a compounding five-level roadmap.
Under the hood
- 01
Smart Document Intake
OCR via Tesseract.js and pdf-parse, with Gemini parsing structured financial data at ~99% accuracy.
- 02
Global Tax Engine
Tax simulation across countries worldwide, projecting net surplus income based on local rules.
- 03
Dynamic Waterfall Roadmap
Five-level compounding goal generator based on net surplus.
- 04
Interactive Goal Workspaces
Per-level workspaces with AI chat, checklists, and progress tracking. 40% higher goal completion.
- 05
LLM Integration Benchmarking
Compared Gemini, OpenRouter, and Groq across 500+ queries for speed, reliability, and cost.
- 06
60fps Canvas Scrollytelling
HTML5 Canvas + Framer Motion, 270 frames synced to device pixel ratio.
HitLab AI
Billboard Hot 100 Predictor
Open case