Hacker News new | ask | show | jobs
by amelius 3254 days ago
Perhaps a crowdsourced approach could help here, where the raw text and relevant images are identified/extracted by the first few readers.