AI Financial Intelligence

An AI-powered financial analysis platform that extracts structured insights and generates standard financial reports directly from raw data — no accounting knowledge required.

Designed for individuals, micro-businesses (warung/toko), and SMEs (UMKM) in Indonesia.

---

Features

Core Analysis

• 3-Statement Generator — automatically produces Income Statement, Balance Sheet, and Cash Flow Statement from uploaded data
• BI Dashboard — revenue breakdown, expense analysis, category drilldown, and profitability charts
• Confidence Review — extraction confidence score with flagged rows before data enters the analysis pipeline

Analysis Modes

Mode	Description
Personal	50/30/20 budgeting framework, health score, spending breakdown, personalized recommendations
Toko/Warung	Simplified view for micro-businesses — busiest day, top-earning categories, restock signals, plain-language tips
Bisnis/UMKM	Full business analysis with 3 statements, BI insight, investment readiness, and benchmark comparison

Multi-Period & Forecasting

• Groups transactions by month automatically
• Linear regression forecast 3 months ahead
• Month-over-month comparison table
• Automatic warnings: declining revenue, cash-tight months, rising costs

Investment Readiness Score (Bisnis/UMKM)

• 0–100 score across 4 dimensions: consistency, profitability, growth, cash flow health
• Grade A–D with strengths and risk indicators
• One-page printable/PDF-exportable summary ready for loan or investor proposals

Benchmark (Bisnis/UMKM)

• Compare your business against Indonesian UMKM sector averages
• Sectors: F&B, Retail/Toko, Jasa/Services, Pertanian/Agribusiness
• Plain-language delta descriptions ("your margin is 8% higher than the sector average")

File Support

• CSV — auto column detection, rule-based parser with AI fallback
• Excel (.xlsx / .xls) — smart sheet and header detection
• PDF — text extraction via PDF.js + structured parsing via Google Gemini AI
• Downloadable templates — Personal and UMKM Excel templates included

---

Tech Stack

Layer	Technology
Framework	React 18 + Vite
Charts	Recharts
Icons	Lucide React
PDF Extraction	pdfjs-dist
Excel Read/Write	SheetJS (xlsx)
CSV Parsing	PapaParse
AI Extraction	Google Gemini API (gemini-2.5-flash)
Styling	Custom CSS with CSS variables (dark/light theme)
State	React useState + useMemo — no external state library
Backend	None — fully client-side

---

Getting Started

Prerequisites

• Node.js 18+
• A free Google Gemini API key (required for PDF parsing only)

Installation

# Clone the repository
git clone <your-repo-url>
cd "AI Financial Intellegence"

# Install dependencies
cd frontend
npm install

Running Locally

cd frontend
npm run dev

Open http://localhost:5173 in your browser.

Environment Variables (optional)

You can pre-load a Gemini API key via environment variable instead of entering it in the UI:

# frontend/.env.local
VITE_GEMINI_API_KEY=your_api_key_here

If not set, the app will prompt you to enter the key through the in-app settings panel. The key is stored in localStorage and never sent anywhere except directly to the Google Gemini API.

Build for Production

cd frontend
npm run build
# Output in frontend/dist/

---

Project Structure

AI Financial Intellegence/
├── frontend/
│   ├── src/
│   │   ├── components/       # React UI components
│   │   │   ├── FileUpload.jsx
│   │   │   ├── ConfidenceReview.jsx
│   │   │   ├── FinancialStatements.jsx
│   │   │   ├── BIDashboard.jsx
│   │   │   ├── MultiPeriodView.jsx
│   │   │   ├── PersonalFinanceView.jsx
│   │   │   ├── MikroView.jsx
│   │   │   ├── InvestmentView.jsx
│   │   │   └── BenchmarkView.jsx
│   │   └── utils/            # Business logic & engines
│   │       ├── parser.js           # CSV/Excel rule-based parser
│   │       ├── pdfParser.js        # PDF extraction + Gemini API
│   │       ├── statementGenerator.js
│   │       ├── multiPeriodEngine.js
│   │       ├── personalFinanceEngine.js
│   │       ├── mikroEngine.js
│   │       ├── investmentEngine.js
│   │       ├── benchmarkEngine.js
│   │       └── templateGenerator.js
│   ├── package.json
│   └── vite.config.js
├── .gitignore
├── v3-roadmap.md
└── README.md

---

Notes

• No backend, no database — all computation happens client-side in the browser. No financial data is stored or transmitted to any server (except to the Gemini API for PDF parsing).
• PDF parsing requires a Gemini API key — the free tier (1,500 requests/day) is sufficient for normal personal use.
• Benchmark data is based on representative Indonesian UMKM statistics (BPS, Bank Indonesia, Kemenkop UKM) and is intended as a general reference, not a formal financial assessment.

---

Status

This project is under active development.

The core platform (V1–V3) is functional. Planned work includes additional data source integrations, improved categorization accuracy, and potential monetization features for power users. See v3-roadmap.md for the full roadmap.

---

License

Private project — not licensed for public distribution.