Okapi, or “What if ripgrep Could Edit?”

Okapi is a command-line tool built on top of ripgrep that enables bulk editing of OCR errors (scannos) across thousands of text files. The author built it while digitizing 150+ years of US Government employee records, where OCR tools like olmOCR still produce many character-level errors. Okapi aggregates regex matches from multiple files into a single editable buffer (inspired by git interactive rebase), allowing multi-select editing in Sublime Text. A companion Rust tool uses Tesseract bounding boxes and fuzzy trigram matching to display the original scanned image inline next to each matched line, providing visual ground truth for ambiguous corrections.