epitran

library1.35.1pypypi✓ verified 85d ago

Epitran is a library for transliterating text written in various writing systems into the International Phonetic Alphabet (IPA). Version 1.35.1 supports 50+ languages and scripts. Release cadence is irregular, with multiple releases in 2025-2026.

data serialization

pip install epitran

github ↗package ↗docs ↗changelog ↗

INSTALL

IMPORT

SIG · EPITRAN

epitran

datapythonv1.35.1

Install

10.8s avg

Import

1767ms

Disk

190MB

Pass rate

8/ 10

Env Coverage8 / 10

glibc

3.9–3.13

musl

3.9–3.13

Install & Compatibility

Where this runs

tested against v1.35.1 · pip install

no network on importno background threads

Install × environment matrix

Each cell = how many times install + import succeeded across repeated harness runs. Partial = flaky.
glibc = Debian/Ubuntu slim · musl = Alpine Linux

musl

glibc

py 3.10

✓ —

✓ 10.3s

py 3.11

✓ —

✓ 9.6s

py 3.12

✓ —

✓ 10.8s

py 3.13

✕ build_error

py 3.9

✓ —

✓ 12.3s

190MB installed

● package 190MB

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

Epitran

✓ from epitran import Epitran

Standard import path.

Initialize Epitran with a language code (e.g., 'eng-Latn') and call .transliterate() to convert text to IPA.

from epitran import Epitran # Create an Epitran object for English (eng-Latn) epi = Epitran('eng-Latn') # Transcribe a word to IPA ipa = epi.transliterate('hello') print(ipa) # Output: həlˈoʊ

Debug

Known issues

gotchaEpitran requires a language code in language-scrīpt format (e.g., 'spa-Latn' for Spanish). Using just 'en' will raise a ValueError.

fix

Use the exact language-script code as specified in the documentation (e.g., 'eng-Latn', 'cmn-Hans').

affects: all

gotchaEpitran uses a local data directory (~/.epitran) for lookup tables. If the data is missing, operations will fail silently or throw errors.

fix

Run `epitran -d` after installation to download and set up the data, or let Epitran auto-download on first import (requires internet).

affects: all

deprecatedThe old `epitran.Epitran()` instantiator may not work with Python < 3.10; as of v1.35.1, Python >=3.10 is required.

fix

Upgrade Python to 3.10 or later. If stuck on older Python, use epitran < 1.35.0.

affects: >=1.35.0

breakingIn v1.34.0, the internal data format changed. Code relying on direct file access to ~/.epitran may break.

fix

Use the public API (Epitran class) only; avoid reading data files directly.

affects: >=1.34.0

Errors

Common errors & fixes

ValueError: Unsupported language: en

Using a plain language code without script (e.g., 'en' instead of 'eng-Latn').

fix

Use full language-script code, e.g., `Epitran('eng-Latn')`. List supported codes: `epitran.list_languages()`.

epitran.Epitran not found

Importing the class incorrectly as `epitran.Epitran` instead of `from epitran import Epitran`.

fix

Import using `from epitran import Epitran`.

FileNotFoundError: [Errno 2] No such file or directory: '.../epitran/data/...'

Epitran data directory (~/.epitran) is missing or incomplete.

fix

Run `epitran -d` on the command line to download all required data. Ensure internet connectivity.

Upgrade

Version history

1.35.1latest on PyPI · released Apr 24, 2026

Audit

Dependencies

unidecoderequiredUsed for transliteration of non-ASCII scripts

Agent activity

4 hits · last 30 days

node

Resources

githubhttps://github.com/dmort27/epitran ↗

packagehttps://pypi.org/project/epitran/ ↗

docshttps://github.com/dmort27/epitran#readme ↗

changeloghttps://github.com/dmort27/epitran/releases ↗