UNICHARSET_EXTRACTOR(1)

💅 UNICHARSET_EXTRACTOR — the unicharset glow-up tool

okay bestie, so this is the unicharset_extractor manual page (section 1 of the manual, if you're keeping track 🎀), and honestly? It's giving "small tool, big responsibility."

---

✨ NAME ✨

unicharset_extractor — reads box files or plain text files and extracts the unicharset from them. That's the whole job. She's focused. She's iconic. 💖

---

🌸 SYNOPSIS 🌸

unicharset_extractor [--output_unicharset filename] [--norm_mode mode] box_or_text_file [...]

And like, what does mode even mean? Bestie, I got you, here's the breakdown:

• 1 = combine graphemes (use this for Latin and other simple scripts) 🥰
• 2 = split graphemes (use this for Indic/Khmer/Myanmar) 🦋
• 3 = pure unicode (use this for Arabic/Hebrew/Thai/Tibetan) ⭐

---

🔥 DESCRIPTION 🔥

So here's the vibe: Tesseract needs to know the set of possible characters it can output. Like, it can't just freestyle — it needs to know what's on the menu. To generate the unicharset data file, you run the unicharset_extractor program on training pages bounding box files, or on a plain text file:

unicharset_extractor fontfile_1.box fontfile_2.box ...

The unicharset will be put into the file ./unicharset if you don't provide an output filename. No filename? She picks one for you. Self-sufficient queen. 👑

NOTE 📌 Use the appropriate norm_mode based on the language. This is not a "pick whichever, it's all the same" situation — it fr fr matters. Pick the right one for your script, bestie.

---

🫶 SEE ALSO 🫶

• tesseract(1)
• unicharset(5)
• https://tesseract-ocr.github.io/tessdoc/Training-Tesseract.html

---

🦋 HISTORY 🦋

unicharset_extractor first appeared in Tesseract 2.00. An elder. A legend. Respect her. 💜

---

🎀 COPYING 🎀

Copyright (C) 2006, Google Inc. Licensed under the Apache License, Version 2.0

---

🌈 AUTHOR 🌈

The Tesseract OCR engine was written by Ray Smith and his research groups at Hewlett Packard (1985-1995) and Google (2006-2018). King behaviour, honestly. 👑

---

and that's the page, bestie — this version dropped on 2026-09-28, and it's UNICHARSET_EXTRACTOR(1) all day, every day. go forth and extract your unicharset ✨💅