Case Study

Handling Real-World PDF Edge Cases

Next.js · Go · PDF Processing · OCR · Workspace Flow

Executive Summary

Focused on the hard parts of document software: files that break assumptions, operations that fail in subtle ways, and workspace states that need to stay recoverable.

The Problem

Large files and multi-step workflows made it easy for users to lose progress.
Different tools needed different file states, which complicated shared routing and workspace logic.
Crop, OCR, and conversion flows could fail depending on the input document structure.

Architecture & Approach

1
Shared upload → workspace → download flow
2
Tool-specific workspace components that isolate logic
3
State sharing between pages to preserve file context
4
Backend validation and guarded processing to fail safely

Challenges & Trade-offs

Upload state had to survive navigation across multiple screens

Crop logic needed careful bounds handling to avoid incorrect output

Different document structures produced inconsistent results in conversion and OCR

Outcomes & Results

More predictable workflows with clearer recovery paths
Better resilience for messy or unusual input files