blyantgrep

Blyantgrep is an open-source command-line tool designed to search through digitized notebooks, ledgers, and other archival material. It extends the traditional grep approach by integrating optical character recognition to convert scanned pages into searchable text and by recognizing pencil annotations that may be faint or ambiguous in scans.

The name blends blyant, the Scandinavian word for pencil, with grep, signaling its goal of handling handwritten and annotated material.

Blyantgrep emerged from a collaborative project among national libraries and university archives, with an initial release in 2022. It targets digitization workflows where handwritten notes accompany typed text.

Design and features include fast keyword search with Unicode support and regular expressions, OCR backends (such as Tesseract) and pencil-mark analysis, layout-aware results, page coordinates, and thumbnail previews. It supports multiple input formats and a plugin system for extensions.

Usage typically involves indexing a collection and performing queries across it. A user might index a directory of scans and search for a term, with Blyantgrep returning matching pages and annotated notes with image coordinates.

Impact and reception: libraries and museums have used blyantgrep to improve access to annotated materials. While OCR accuracy for historical handwriting remains imperfect, the tool complements human review and accelerates archival research.