Anki Assistant

Anki Assistant is a Node.js application that helps you create and update Anki decks for language learning. It allows you to input words or phrases in your native language, translates them to your target language, and adds them as cards to an Anki deck.

It uses Anthropic for fluent translations and Narakeet for audio generation.

Installation

1. Clone this repository:

   git clone https://github.com/yourusername/anki-assistant.git
   cd anki-assistant

2. Install the required dependencies:

   npm i

3. Set up the environment variables:

   npm run setup

4. Open the .env file and update the following environment variables:

   ANTHROPIC_API_KEY=your_anthropic_api_key
   NARAKEET_API_KEY=your_narakeet_api_key
   
   NATIVE_LANGUAGE=your_native_language # The language you speak
   TARGET_LANGUAGE=your_target_language # The language you want to learn
   
   NATIVE_LANGUAGE_VOICE=your_native_language_voice # The voice to use for the native language audio
   TARGET_LANGUAGE_VOICE=your_target_language_voice # The voice to use for the target language audio

   Replace the placeholder values with your actual API keys and preferences.

   You will need API keys for both Anthropic and Narakeet.

   • Link to Anthropic: https://console.anthropic.com/dashboard
   • Link to Narakeet: https://www.narakeet.com/
   • Link to Narakeet voices: https://www.narakeet.com/languages/

Usage

Interactive Mode

1. Run the application:

   npm start

2. Follow the prompts in the console:

   • Enter a word or phrase in your native language.
   • The application will provide a translation.
   • Confirm if the translation is correct (y/n).
   • Repeat this process for as many words as you want to add.
   • Enter 'q' when you're done adding words.

3. The application will create an Anki deck file (.apkg) in the project directory.

4. Import the generated .apkg file into Anki to see your new cards.

Batch Mode

In addition to the interactive mode, you can process multiple phrases at once using batch mode:

1. Create a phrases.txt file in the project directory (you can copy from phrases.txt.example):

   cp phrases.txt.example phrases.txt

   Then edit phrases.txt with your phrases, one per line:

   # Lines starting with # are ignored as comments
   hello
   goodbye
   thank you
   how are you

2. Run the batch processor:

   npm run batch

3. For each phrase, the application will:

   • Show the translation
   • Ask for confirmation
   • Generate audio files
   • Add the card to your deck if confirmed

This is particularly useful when you have a list of phrases prepared in advance.

Audio Tapes

You can also create audio learning tapes that combine multiple phrases into a single MP3 file. Each phrase follows this pattern:

1. Native language phrase
2. 5-second pause
3. Target language phrase (normal speed)
4. 3-second pause
5. Target language phrase (0.7x speed)
6. 3-second pause
7. Target language phrase (0.85x speed)
8. 3-second pause
9. Target language phrase (normal speed)
10. 3-second pause before next phrase

This format is particularly useful for listening and pronunciation practice, when you may not be able to read the text.

1. Create an audio-phrases.txt file in the project directory (you can copy from audio-phrases.txt.example):

   cp audio-phrases.txt.example audio-phrases.txt

   Then edit audio-phrases.txt with your phrases. You can either:

   1. Add just the native language phrase:

      I usually eat breakfast at home
      She can speak three languages very well
      The cat likes to sleep under the table

   2. Or include both native phrase and translation, separated by ---:

      I usually eat breakfast at home --- Es parasti ēdu brokastis mājās
      She can speak three languages very well --- Viņa var ļoti labi runāt trīs valodās
      The cat likes to sleep under the table --- Kaķim patīk gulēt zem galda

   If you provide translations, they will be used directly. Otherwise, the application will generate translations automatically.

2. Run the audio tape creator:

   npm run audio

3. The application will:

   • Translate each phrase
   • Generate audio for both languages
   • Create slower versions for practice
   • Combine everything into a single MP3 file

The output will be saved as [TARGET_LANGUAGE]_audio_tape.mp3 in your project directory.

For more information and the source code, visit the GitHub repository.