PDFy

ADR-0001: Website-First Architecture With Python Analysis Backend

Status

Accepted

Context

The product needs to start as a website-style PDF scanning experience and later expand without replacing the analysis core. Deep PDF malware analysis also aligns much better with Python-based tooling than a Node-only stack.

Decision

Use:

• Next.js for the website and user-facing API layer
• Python services for PDF analysis
• queued background workers for advanced analysis and cleanup
• short-lived storage with immediate deletion as the default retention mode

Consequences

• The product can evolve into a larger website platform without rewriting the analyzer.
• The analysis engine can use stronger PDF security tooling.
• Deployment is more complex than a single-process app, but the architecture matches long-term product needs.