BrailleKit

Braille Translation for Everyone

This changelog lists the changes that affect what you get when you translate text, music, or math into braille and back.

[Unreleased]

Added

  • Music braille now supports breve, longa, and 256th notes (and their rests) in MusicXML scores; previously these note values were dropped.

Fixed

  • Back-translation of phrase contractions now works correctly. Multi-word phrase signs (for example the French au contraire) are turned back into the full phrase instead of breaking apart into separate characters.
  • A contracted table that builds on a base table now correctly keeps the base table's settings (such as script handling, math routing, music signs, and hyphenation) when it doesn't override them. Previously some of these could be lost.

[0.8.0] — 2026-06-03

Added

  • Automatic updates on Windows. BrailleKit now checks for updates daily in the background and offers signed installer updates through a standard "Software Update" dialog. A "Check for Updates…" entry is also available in the Help menu.
  • Spanish Grade 2 (es-g2) now provides full coverage of the official Spanish contracted braille standard (CBE B-16-1, V3 2025). The table grew from 128 to 749 word contractions plus 2 phrase contractions, and is now marked stable.
  • Portuguese Grade 2 (pt-g2) now provides full coverage of the authoritative Brazilian standard (IBC Estenografia Braille 2006, Section I). The table grew from 81 to 149 word contractions and is now marked stable.
  • Braille music from MusicXML (alpha): selecting a MusicXML score now produces international braille music notation following the New International Manual of Braille Music Notation (World Blind Union, 1996). The note, rhythm, and interval signs are language-independent; the score's text (title, composer, part names, lyrics, directions) is still translated through your selected language table. This first version covers notes, rests, octave marks, dotted notes, accidentals, key and time signatures, bar lines, and chords. Precise lyric-to-note alignment for vocal scores is planned for a later release.

Fixed

  • Spanish Grade 2 (es-g2): removed several broken entries and corrected 16 contractions that conflicted with the standard (including alguno/alguna and related forms, the ninguno and otra families, question words, and others).
  • Portuguese Grade 2 (pt-g2): corrected 5 contractions that did not match the IBC standard (jamais, você, qual, quem, tudo).

[0.7.0] — 2026-04-29

Added

  • Math inside prose is now translated automatically. When a language table is set up for it (for example UEB Grade 2 paired with Nemeth), text containing math (MathML) sections is detected and each math section is translated through the math translator, with the correct switch indicators inserted at every text-to-math boundary.
  • Cross-word phrase contractions. Tables can now map a multi-word phrase to a single braille output (for example the French il y a), so common phrases translate as one unit.

Changed

  • English Nemeth (en-nemeth): the numeric indicator is now emitted in the positions the BANA Nemeth Code requires (at the start of an expression and after spaces). This changes the braille for every expression that starts with a digit; the output now matches the standard. Other math tables are unaffected.

[0.6.2] — 2026-04-21

Added

  • Portable Word (DOCX) output with an embedded braille font. A new option embeds a public-domain braille font inside the .docx file, so the document displays braille correctly even on computers that don't have a braille font installed. This adds about 50 KB to the file.
  • Rich Text (RTF) output, with the same embedded-braille-font option. Note: the macOS text engine (TextEdit, textutil) does not support embedded fonts in RTF, so use DOCX on those readers.
  • The Document tab in the app now offers Rich Text (.rtf) as an output format alongside DOCX, plus a font option "Braille CC0 (embedded, portable)" that turns on font embedding for whichever format you choose. Save dialogs and accessibility labels (English and German) were extended to match.
  • Afrikaans Grade 2 (af-g2): new alpha table covering South African UBC-af contractions.
  • Swahili Grade 2 (sw-g2): new alpha table covering roughly 80% of typical Swahili Grade 2 usage.

Changed

  • English UEB Grade 2 (en-ueb-g2): improved handling of ambiguous letter sequences that need a grade-1 indicator, so words like Al's, Blvd, CD, and alt. now translate correctly.
  • English UEB Grade 2 (en-ueb-g2): shortforms like blind, braille, first, friend, good, great, and letter now also fire at the start of a capitalised compound word (for example Blindcraft).
  • English UEB Grade 2 (en-ueb-g2): 21 proper nouns (such as Boone, Chisholm, Esther, Jamestown, Hades) no longer get incorrect contractions applied inside the word.

Fixed

  • Command-line --version and --help were previously treated as text and translated into braille; they now print version and usage information as expected.
  • French Grade 2 (fr-g2): removed extraction-noise entries and restored 58 French shortforms to their correct two-cell forms.

[0.6.1] — 2026-04-20

Changed

  • Grade 2 number mode now stays active across punctuation listed for a table (by default comma and period), so a run of digits separated by such punctuation is kept in one number segment. Previously Grade 2 reset number mode too early. Affects Norwegian, German, French, Italian, and others.
  • English UEB Grade 2 (en-ueb-g2): contractions like in and enough now extend correctly before apostrophe endings ('s, 't, 'd, 'll, 've, 're).
  • The two-letter strong signs (sh, th, ch, st, wh, gh, ar, ou) are no longer used when they stand as an entire word on their own (for example sh now translates as two letters, not the single contraction).

Fixed

  • English UEB Grade 2 (en-ueb-g2): added apostrophe-ending support across all alphabetic wordsigns and shortforms; fixed con and dis exception handling; stopped ing from firing at the start of a word after a hyphen; and removed several bogus shortforms that came from earlier extraction errors.
  • German Grade 2 (de-g2): fixed several remaining mismatches against the German standard (including ihrige, lässt's, möcht's, Man).
  • Norwegian Grade 1 (no-g1): added 10 missing symbol mappings (&, §, ©, π, °, |, =, +, @, _, $) and corrected the double-quote mapping.
  • French Grade 2 (fr-g2): marked 20 outdated entries as rejected; both BrailleKit and the reference tool now emit the current-edition shapes.

[0.6.0] — 2026-04-15

Added

  • Redesigned desktop app: a guided wizard flow for text and document conversion, eBraille export, EPUB and Markdown import, drag-and-drop for single files and batches, live preview, system-aware dark mode, and a table maturity indicator (stable / beta / alpha) in the table picker.
  • Grade 2 back-translation (braille back to text) using the table's contractions in reverse.
  • Capital passage indicator support.
  • German compound-word handling improved (recursive boundary detection), correctly translating 77 more words.
  • New stable language coverage and additions, including Norwegian, Swedish, English UEB Grade 2, and French Grade 2 entries; Welsh Grade 2 (cy-g2) and Portuguese Grade 2 (pt-g2) promoted to stable.

Changed

  • Table accuracy improvements on major tables, including German Grade 2 (de-g2) coverage of the German standard rising from 90% to 98%, English UEB Grade 2 fixes from the 2024 UEB Rules, and added French abbreviations.

Fixed

  • Korean Hangul: fixed a character-encoding problem that could corrupt Korean characters on Windows builds.
  • German Grade 2 (de-g2): fixed a compound-boundary issue that wrongly blocked the st contraction.
  • Capital indicator now suppressed correctly after hyphens in compounds and before the letter-prefix indicator.
  • English UEB Grade 2 (en-ueb-g2): fixed con- false positives in words like console and contest; added missing Beatrice, Beatrix, Belinda, and acknowledge shortforms.
  • Command-line input piped from another program now works correctly in Grade 2 mode.

[0.5.0] — 2026-03-24

First versioned release: 49 braille tables across more than 40 languages.

Added

  • Chinese braille: Chinese characters are converted through pinyin using a built-in dictionary; Chinese Grade 1 (zh-g1) accuracy rose from 0% to 91.8%.
  • A new feature for non-Latin scripts that emits a single indicator before a run of Latin text, raising Russian Grade 1 (ru-g1) accuracy from 94.0% to 98.8%.
  • Japanese (Tenji) overhaul: added all Hiragana and Katakana, fixed a わ/を swap, and added many combination characters; Japanese Grade 1 (ja-g1) accuracy rose from 0% to 97.3%.

Changed

  • Accuracy improvements across more than 20 tables, with large gains for Spanish Grade 2 (15.0% → 70.2%), Hindi Grade 1 (77.8% → 98.0%), Portuguese Grade 2 (40.3% → 52.1%), and French Grade 2 (28.9% → 40.7%), plus smaller gains for Arabic, Dutch, Russian, Czech, Polish, and others.
  • Root causes fixed included character-mapping errors (Polish ż, Hindi फ, Spanish é, Arabic shadda position), missing foreign and loanword characters across many tables, and missing Grade 2 contractions for Spanish, Portuguese, and French.

Fixed

  • Chinese braille: fixed loading of the Chinese initials and finals so they are correctly available during translation.