NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #655 most downloaded on PyPI
ASCII transliterations of Unicode text
Last release 1 years ago
24 Apr 2025
Release timing varies
gaps range from 9 days to 1.5 years
Nearly every release is documented
notes for 29 of 29 stable releases
3 versions withdrawn
withdrawn after publishing
17 years old
32 releases · first in 2009
Add replacements for playing card suits and chess pieces (thanks to Alan Davidson)
Fix replacement for U+1E9E "LATIN CAPITAL LETTER SHARP S" (thanks to Thomas Voss)
One column per quarter.
Add missing replacements for katakana punctuation (thanks to Emil Hammarberg)
* No changes. Re-upload to PyPi.
Remove trailing space in replacements for vulgar fractions.
Add some missing replacements for symbols in the U+21xx and U+1F1xx pages (thanks to @cheznewa on GitHub)
Command-line utility now reads input line-by-line, making it usable with large files (thanks to Jan-Thorsten Peter)
Re-upload because PyPI was missing Requires-Python metadata for the .tar.gz package.
Fix issue with wheel package falsely claiming support for Python 2.
Improvements to Hebrew and Yiddish transliterations (thanks to Alon Bar-Lev and @eyaler on GitHub)
Add 'errors' argument that specifies how characters with unknown replacements are handled. Default is 'ignore' to replicate the behavior of older vers
Add some missing replacements in the U+23xx page.
Fix tests failing on PyPy 7.1.1
Add more Latin letter variants in U+1F1xx page.
Improve transliteration of Hebrew letters (thanks to Alon Bar-Lev)
Move to semantic version numbering, no longer following version numbers from the original Perl module. This fixes an issue with setuptools (>= 8) and
Add U+2116 NUMERO SIGN (thanks to Alan Davidson)
Fixed transliteration of circled Latin letters and numbers (thanks to Krzysztof Jurewicz).
Add unidecode_expect_ascii() and unidecode_expect_nonascii() functions for performance critical applications. (thanks to Israel Saeta Perez)
Add a command line utility. (thanks to Andrew Udvare)
Add some missing Latin letter variants used by popular "Unicode text converter" tools.
* Corrected ChangeLog.
Revert transliteration for Latin characters with umlauts back to language-neutral, unaccented characters.
Use German transliteration for Latin characters with umlauts. (thanks to Geoffrey Teale)
Make argument type warning refer to unidecode()'s caller.
Issue a warning on Python 2.x if an object that is not an instance of unicode class is passed to unidecode().
* Added trove classifiers.
Fixed Python 2.5 installation issue.
License change to GPL to fit better with other Python modules and to remove the issue of depending on Perl for licensing terms.
Unit test fixes for Python builds using 16-bit Unicode representation (see PEP 261)
Support for a lot of various less common variants of Latin letters and digraphs.
* First public release
Your coding agent can read these notes before it upgrades. Set up the MCP server →