🏛️ ADR-001: FastAPI for Backend Services

Context

We need a modern, high-performance web framework to build the SplitDee backend. The backend must handle user requests, real-time message notifications, and image upload streams for AI slip OCR processing.

Alternatives Considered

1. Django / Django REST Framework (DRF): Highly robust but carries significant overhead, lacks native asynchronous performance out of the box, and setup is complex.
2. Node.js (Express / NestJS): Excellent for real-time and I/O tasks, but limits integration with the Python AI ecosystem (OCR, image matching libraries).
3. FastAPI (Selected): A modern, high-performance Python framework based on standard type hints.

Decision

We selected FastAPI for our backend microservices.

Consequences

• Asynchronous Execution: FastAPI’s ASGI support permits high-concurrency requests, which is essential for asynchronous slip scanning and chat message streams.
• Python Ecosystem: Native integration with Python data science, OCR, and AI libraries.
• Auto-generated Documentation: Automatic OpenAPI and Swagger UI generation streamlines frontend integration.
• Strict Typing: Built on Pydantic, ensuring rigorous request/response data validation.