Hacker News Reader

OCR is a solved problem. It's like one call to tesseract and you're done.

The fact they haven't done this just shows how inept their employees are.

Tesseract has a horrific API and decades of technical debt. Add to this, it's also very sensitive to fonts. So... yeah, OCR is 'solved', but that doesn't mean it's easy.

Solved for most the image text I see on social.

Reply on news.ycombinator.com