The aim of this project is collaborating in the creation of high quality, free and open source Anki decks for language learning. Check the website for more details AnkiLangs.org.
AnkiLangs is not part of nor necessarily endorsed by Anki (https://apps.ankiweb.net/).
This is a monorepo containing everything.
Example card to learn Spanish as an English speaker. Question side and answer side.
Find more screenshots on the website.
See here for instructions and to download decks to try yourself.
Both code contributions and audio from (as good as) native speakers are highly appreciated! Have a look at our instructions for contributors.
- Python 3 (Installation).
- uv (Installation).
- Anki (Installation).
- Within Anki the CrowdAnki add-on (code 1788670778). Add-on installation.
git clone https://github.com/ankilangs/ankilangs
cd ankilangs/If you want you can build the decks (i.e. convert the CSV files into Anki decks). Note that you do not need to do this in order to make a contribution. If you want to improve a deck you can stick to the "Contribute changes" section above and leave the complicated stuff to us 🙂.
uv run al-tools generate -i src/data/
uv run al-tools check -i src/data/
uv run brainbrew run recipes/source_to_anki_minimal_pairs.yaml
uv run brainbrew run recipes/source_to_anki_625_words.yamlOpen Anki and via File / CrowdAnki: Import from disk import any of the build/ subdirectories of this
Git repository.
Then you may review them like any deck.
To maintain code quality, this project uses Ruff for both linting and formatting Python code.
To check for linting issues:
uv run ruff check .To automatically format all Python files:
uv run ruff format .This project uses pytest for testing. To run the test suite:
uv run pytestThe tests are located in the tests/ directory and cover core functionality including:
- Ambiguity detection algorithms
- Data processing and validation
- CSV file handling and transformations
- Copy and adapt one of the directories in
src/note_models/
# Probably leave the following unchanged
export AL_SRC_NAME="en_to_pt"
export AL_SRC_NAME_2="EN to PT"
export AL_SRC_LANG_NAME="Portuguese"
export AL_SRC_LISTENING="Listening"
export AL_SRC_PRONUNCIATION="Pronunciation"
export AL_SRC_READING="Reading"
export AL_SRC_SPELLING="Spelling"
# CHANGE ME!
export AL_DST_NAME="de_to_fr"
export AL_DST_NAME_2="DE to FR"
export AL_DST_LANG_NAME="Französisch"
export AL_DST_LISTENING="Listening"
export AL_DST_PRONUNCIATION="Pronunciation"
export AL_DST_READING="Reading"
export AL_DST_SPELLING="Spelling"
## Use the following if the source language is not EN
#export AL_DST_LISTENING="Hörverständnis"
#export AL_DST_PRONUNCIATION="Aussprache"
#export AL_DST_READING="Leseverständnis"
#export AL_DST_SPELLING="Rechtschreibung"
cp "src/headers/description_${AL_SRC_NAME}-625_words.html" \
"src/headers/description_${AL_DST_NAME}-625_words.html"
cp -r "src/note_models/vocabulary_${AL_SRC_NAME}" \
"src/note_models/vocabulary_${AL_DST_NAME}"
sed -i "s/id: .*/id: `uv run python -c "import uuid; print(uuid.uuid4())"`/" \
"src/note_models/vocabulary_${AL_DST_NAME}/note.yaml"
find "src/note_models/vocabulary_${AL_DST_NAME}/" -type f \
-exec sed -i "s/${AL_SRC_NAME}/${AL_DST_NAME}/g" {} +
find "src/note_models/vocabulary_${AL_DST_NAME}/" -type f \
-exec sed -i "s/${AL_SRC_NAME_2}/${AL_DST_NAME_2}/g" {} +
find "src/note_models/vocabulary_${AL_DST_NAME}/" -type f \
-exec sed -i "s/${AL_SRC_LANG_NAME}/${AL_DST_LANG_NAME}/g" {} +
find "src/note_models/vocabulary_${AL_DST_NAME}/" -type f \
-exec sed -i "s/| ${AL_SRC_LISTENING}/| ${AL_DST_LISTENING}/g" {} +
find "src/note_models/vocabulary_${AL_DST_NAME}/" -type f \
-exec sed -i "s/| ${AL_SRC_PRONUNCIATION}/| ${AL_DST_PRONUNCIATION}/g" {} +
find "src/note_models/vocabulary_${AL_DST_NAME}/" -type f \
-exec sed -i "s/| ${AL_SRC_READING}/| ${AL_DST_READING}/g" {} +
find "src/note_models/vocabulary_${AL_DST_NAME}/" -type f \
-exec sed -i "s/| ${AL_SRC_SPELLING}/| ${AL_DST_SPELLING}/g" {} +Edit the deck description file:
vim "src/headers/description_${AL_DST_NAME}-625_words.html"To generate new UUIDs you can use this command:
uv run python -c "import uuid; print(uuid.uuid4())"Edit recipes/source_to_anki_625_words.yaml:
- Add the new source files under
generate_guids_in_csvs - Copy a block under
note_models_from_yaml_partand:- Edit the
part_idandfile
- Edit the
- In the
headers_from_yaml_partsection copy a block and:- Edit the
part_id - Edit the
name - Replace
crowdanki_uuidwith a new UUID - Set the path of the
deck_description_html_file
- Edit the
- In the
notes_from_csvssection copy a block and:- Adapt all strings to the new language
- Copy a
generate_crowd_ankisection and replace all strings with the new language
Create the new CSV files you need under src/data/ by copying and adapting
existing files. Note that the guid columns must stay empty since they have
to be unique and will automatically be generated during build.
You need a Google Cloud account and need to be authenticated.
Then execute, for example:
uv run al-tools audio -i src/data/625_words-base-de_de.csv -o src/media/audio/de_DE/
uv run al-tools audio -i src/data/625_words-base-pt_pt.csv -o src/media/audio/pt_PT/
uv run al-tools audio -i src/data/625_words-base-it_it.csv -o src/media/audio/it_IT/
uv run al-tools audio -i src/data/625_words-base-fr_fr.csv -o src/media/audio/fr_FR/
uv run al-tools audio -i src/data/625_words-base-en_us.csv -o src/media/audio/en_US/
uv run al-tools audio -i src/data/625_words-base-es_es.csv -o src/media/audio/es_ES/To release a deck do the following:
- Update the version in the
src/headers/description*file - Build as described above
- Update the version in the
src/headers/description*file to the next dev version - Import the deck into Anki
- Export from Anki:
- Include media
- Support older Anki versions
- Rename the exported deck so it ends with the version e.g.
- 0.0.1 - Add a new Git tag
- Create a release on GitHub with the exported deck
- Update the README and website with download links