Architecture Overview

odirscan is currently a single-process application with a clear separation of concerns across packages. The architecture is designed for future evolution toward a distributed system.

Data Flow

Component Responsibilities

Component	Package	Role
Scanner	internal	Crawls directory listings and tags findings with HTTP metadata
Mullvad	internal/mullvad	Manages relay list, proxy selection, and VPN connectivity checks
Config	config	Holds all configuration structs with default constructors
Messages	messages	Defines the request/response types exchanged between components
Model	model	Data types for scan findings and results
Web	web	HTTP server with embedded SPA for browsing results

Design Decisions

• colly for crawling -- Provides async collection, proxy switching, and request throttling out of the box.
• Actor-like messages -- ScanRequestMessage, ScanResultMessage, and ScanFindingMessage decouple the caller from the scanner implementation, preparing for future RPC boundaries.
• Embedded web UI -- The HTML dashboard is compiled into the binary via go:embed, eliminating runtime file dependencies.