CREDIT_IQ is a professional credit intelligence platform designed to automate the credit appraisal lifecycle for corporate borrowers. The system leverages artificial intelligence to ingest, classify, and cross-validate financial data from multiple sources, providing an auditable and explainable risk score.
The system initiates the appraisal process by capturing core entity details (CIN, PAN, Sector) and specific loan requirements. This creates a centralized record for all downstream analysis.
Automated classification of financial documents (GST returns, Bank Statements, ITRs, etc.) using semantic similarity models. A human-in-the-loop interface ensures classification accuracy before data extraction.
A specialized engine cross-references data points across disparate sources to identify inconsistencies. It flags discrepancies such as revenue mismatches between GST filings and bank records to mitigate fraud risk.
High-precision extraction of semi-structured data from PDFs and spreadsheets. The system maps extracted fields to a standardized schema, allowing for consistent financial analysis across different document formats.
Synthesizes primary data and external intelligence into a comprehensive credit memo.
- Risk Scoring: Utilizes gradient boosting models with SHAP-based explainability.
- SWOT Analysis: An AI agent generates structured insights based on document content and financial findings.
- Reporting: Automated generation of a professional PDF Credit Intelliigence Report.
| Component | Technology |
|---|---|
| Backend | FastAPI, SQLAlchemy, Pydantic |
| Frontend | React, Vite, TailwindCSS |
| NLP & AI | LangChain, Anthropic Claude, OpenAI gpt-4o |
| Data Processing | Pandas, PyMuPDF, Tesseract OCR |
| Scoring | XGBoost, SHAP |
| Reporting | ReportLab |
- Node.js (v18+)
- Python (v3.9+)
- PostgreSQL
- Navigate to the
backenddirectory. - Install dependencies:
pip install -r requirements.txt
- Configure environment variables in a
.envfile based on.env.example. - Run the application:
python -m app.main
- Navigate to the root directory.
- Install dependencies:
npm install
- Start the development server:
npm run dev
The project includes a docker-compose.yml for containerized deployment:
docker-compose up --buildbackend/: FastAPI application, database models, and AI services.src/: React frontend source code.docs/: Supplemental project documentation.uploads/: Local storage for ingested documents (git-ignored).
For more detailed information, please refer to the files in the docs/ directory: