Skip to main content

​
System Architecture

CreditNexus follows a layered architecture with clear separation of concerns:

​
Backend Layer

  • API Layer: FastAPI with dependency injection and Pydantic validation
  • Service Layer: Business logic and orchestration
  • Data Layer: SQLAlchemy 2.0 with PostgreSQL
  • Policy Layer: Policy-as-code engine for compliance enforcement

​
Frontend Layer

  • React 18: Modern React with hooks and functional components
  • TypeScript: Type-safe development
  • Tailwind CSS: Utility-first styling
  • FDC3 Integration: Desktop interoperability

​
AI/ML Layer

  • LLM Abstraction: Support for multiple providers (OpenAI, vLLM, HuggingFace)
  • Extraction Chains: Simple and map-reduce extraction strategies
  • Satellite Analysis: TorchGeo for NDVI verification

​
Data Flow

  1. Document Upload: User uploads credit agreement
  2. Text Extraction: Document text is extracted
  3. AI Extraction: LLM extracts structured CDM data
  4. Policy Validation: Policy engine validates compliance
  5. CDM Event Creation: Policy decisions stored as CDM events
  6. Storage: Data persisted in PostgreSQL with audit logging

​
Key Principles

  • CDM Compliance: All data follows FINOS CDM standards
  • Policy-as-Code: Deterministic policy enforcement
  • Type Safety: Full TypeScript and Pydantic validation
  • Audit Trail: Complete audit logging for compliance

​
Platform Scale & Capabilities

CreditNexus is built with enterprise-scale architecture and comprehensive feature coverage:

​
Component Statistics

  • 215+ Prompt Templates - AI prompts across 16 files covering LMA-compliant clause generation, industry-specific templates, and multi-scenario support
  • 11 Specialized Agents - AI agents for document analysis, satellite verification, classification, research, and workflow automation
  • 121 Policy Rules - Regulatory compliance rules across 21 files covering MiCA, Basel III, FATF, ESG, and green finance
  • 94 Tool Modules - Service modules, utilities, and LangChain integration chains for comprehensive functionality
  • 76 Data Models - Pydantic models ensuring type safety and FINOS CDM compliance across all data structures
  • 54 Database Tables - SQLAlchemy 2.0 models covering users, documents, workflows, deals, payments, and analytics
  • 147 API Endpoints - RESTful endpoints providing comprehensive programmatic access to all platform features
  • 100% CDM Compliant - Full FINOS Common Domain Model compliance ensuring interoperability with financial systems

​
Platform Architecture Highlights

AI & Automation:
  • 215+ prompt templates for consistent AI behavior
  • 11 specialized agents for complex workflows
  • Multi-provider LLM support (OpenAI, vLLM, HuggingFace)
  • 20 LangChain integration chains
Compliance & Policy:
  • 121 policy rules across 8 categories
  • Real-time policy enforcement engine
  • CDM-compliant event generation
  • Three-tier decision system (ALLOW/BLOCK/FLAG)
Data & Integration:
  • 76 Pydantic models for type safety
  • 54 database tables with SQLAlchemy 2.0
  • 147 API endpoints for programmatic access
  • Full CDM standardization
For detailed component breakdown and capabilities assessment, see the Platform Statistics documentation.