Analysis Dashboard

SHARAH - Instant Shariah Compliance Validation

Validate Islamic financial products in seconds, not weeks.

Inspiration

The Islamic finance industry faces a critical bottleneck: validating new financial products for Shariah compliance typically takes 3-5 weeks when reviewed by Shariah scholars. This lengthy process:

Slows down product launches
Increases costs for Islamic financial institutions
Creates friction for fintech innovation in the Muslim world

We recognized an opportunity to leverage AI and rule-based systems to provide instant preliminary compliance validation—helping banks and fintech companies accelerate their product development while maintaining Islamic financial principles.

What it does

SHARAH is a Shariah compliance validation API for Islamic financial products. Users upload PDF documents or submit product specifications, and SHARAH analyzes them against Islamic finance principles in real-time.

Key Features

Feature	Description
Instant Analysis	Validates products in <1 second
Compliance Verdict	Clear result: Compliant / Non-Compliant / Uncertain
Confidence Score	0-100% confidence in the assessment
Detailed Reasoning	Explains why a product is Halal or Haram
Issue Detection	Identifies violations like Riba, Gharar, problematic late fees
Evidence Citations	References Shariah principles and fatwas used
Recommendations	Actionable next steps for redesign or scholar review

How we built it

Architecture

┌─────────────────────────────────────────────────────────┐
│                   Frontend (React/Vite)                  │
│         - PDF Upload (drag & drop)                       │
│         - Results Display with confidence bars           │
│         - Modern dark theme UI                           │
└────────────────────────┬────────────────────────────────┘
                         │ POST /api/shariah-check
                         ▼
┌─────────────────────────────────────────────────────────┐
│                 Backend (FastAPI/Python)                 │
│  ┌───────────────────────────────────────────────────┐  │
│  │              Shariah Engine                        │  │
│  │  - Feature Extraction (keyword detection)         │  │
│  │  - Contract Classification                        │  │
│  │  - Compliance Checks (Riba, Gharar, etc.)        │  │
│  │  - Confidence Scoring                             │  │
│  └───────────────────────────────────────────────────┘  │
│  ┌───────────────────────────────────────────────────┐  │
│  │              RAG System                            │  │
│  │  - Knowledge Base (12 Shariah chunks)             │  │
│  │  - Retriever (keyword matching)                   │  │
│  │  - Reasoner (evidence citation)                   │  │
│  └───────────────────────────────────────────────────┘  │
└─────────────────────────────────────────────────────────┘

Tech Stack

Backend:

Python 3.13+ / FastAPI / Uvicorn
Custom RAG implementation

Frontend:

React + Vite
TailwindCSS
Drag-and-drop PDF upload

Knowledge Base:

12 Markdown documents covering Islamic finance principles
Sourced from AAOIFI standards, Quranic verses, IFSB guidelines

Shariah Principles Covered

Principle	Description
Riba	Interest prohibition on loans
Gharar	Excessive uncertainty in contracts
Murabaha	Cost-plus sale (permitted)
Musharaka	Equity partnership (permitted)
Mudaraba	Profit-sharing partnership (permitted)
Ta'widh	Late fee regulations
Maysir	Gambling/speculation prohibition

Challenges we ran into

1. Encoding Complex Shariah Rules

Translating nuanced Islamic finance principles into deterministic rule-based logic while maintaining accuracy. Many Shariah rulings have edge cases requiring scholarly interpretation.

2. Contract Classification

Distinguishing between:

Loan-like structures (Haram if interest-based)
Legitimate profit-sharing (Halal)
Asset-backed sales (Halal if properly structured)

3. Confidence Calibration

Determining appropriate confidence scores that reflect both:

Clarity of the product structure
Severity of potential violations

4. Handling Ambiguity

Products with vague terms (Gharar) or missing information needed special handling to avoid false positives/negatives.

5. Building the Knowledge Base

Curating accurate, cited Shariah rulings from authoritative sources:

AAOIFI Shariah Standards
Quranic verses (2:275-276)
OIC Fiqh Academy rulings
IFSB Core Principles

Accomplishments that we're proud of

Functional Compliance Engine — Correctly identifies Riba violations, Gharar risks, and partnership compliance issues
Comprehensive Knowledge Base — detailed source data covering major Islamic finance principles with proper citations
RAG Integration — Evidence-backed reasoning in every compliance assessment
Clean Architecture — Proper separation of concerns (API → Service → Engine → Data layers)
Modern UI/UX — Intuitive drag-and-drop interface with real-time feedback
Speed — ~5 second response times for compliance validation

What we learned

Islamic Finance Fundamentals

Riba: Any premium charged on money loans is strictly Haram
Gharar: Excessive uncertainty invalidates contracts
Musharaka/Mudaraba: Profit-and-loss sharing partnerships are permitted
Murabaha: Cost-plus sales are permitted when properly structured

Technical Insights

Rule-based vs ML: For compliance systems, deterministic logic provides transparency and auditability that black-box ML cannot match
RAG Architecture: Effective retrieval with clear upgrade paths
Domain Knowledge Encoding: Authoritative sources (AAOIFI, IFSB) are critical for domain-specific AI

What's next for SHARAH

Priority	Feature	Description
1	LLM Integration	Semantic embeddings for nuanced document analysis
2	Enhanced PDF Parsing	Better OCR and structured extraction
3	Arabic Support	Process original Shariah texts
4	Scholar Dashboard	Review workflow for Shariah scholars
7	Continuous Learning	Scholar feedback improves the engine