Intelli-Credit (Cappri)

Intelli-Credit (Cappri) is a modular, AI-driven platform designed to automate and streamline the corporate lending process. By leveraging specialized micro-agents and advanced OCR capabilities, Cappri transforms manual financial assessments into a transparent and explainable decision-making pipeline.

Key Features

• Automated Document Ingestion: Processes complex financial documents, including GST filings, bank statements, and corporate records.
• Intelligent OCR Pipeline: Utilizes a hybrid approach featuring Chandra OCR for local processing alongside specialized APIs for high-accuracy data extraction.
• Modular Micro-Agent Architecture:
  • Fraud Detection: Identifies anomalies and forensic inconsistencies in financial data.
  • Capacity Calculation: Evaluates debt-servicing capabilities and financial health.
  • Forensic Reconciliation: Automatically reconciles data across disparate sources to prevent manual bottlenecks.
• Explainable AI: Provides transparent credit decisions with clear justifications, ensuring compliance and trust in the appraisal process.
• FastAPI Backend: A high-performance, asynchronous Python-based API layer.

Architecture

The system is built on a modular pipeline architecture, allowing for independent scaling and updates of specific appraisal components.

1. Ingestion Layer: Captures and digitizes raw financial documents.
2. Analysis Layer: Specialized micro-agents process extracted data for risk, fraud, and capacity.
3. Decision Engine: Aggregates agent outputs to provide a comprehensive credit score and recommendation.
4. API Layer: Serves results via a RESTful interface.

Tech Stack

• Language: Python
• Framework: FastAPI
• OCR: Chandra OCR / Specialized Extraction APIs
• Architecture: RAG (Retrieval-Augmented Generation)
• Containerization: Docker

Installation

1. Clone the repository:

   git clone https://github.com/roshanRobyn/IntCredit.git
   cd IntCredit

2. Set up a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Configuration: Create a .env file in the root directory and add necessary API keys and environment variables:

   OCR_API_KEY=your_key_here
   DATABASE_URL=your_db_connection_string

Usage

To start the FastAPI server:

uvicorn main:app --reload

The API documentation will be available at http://127.0.0.1:8000/docs.

License

Distributed under the MIT License. See LICENSE for more information.