SRT/CSV to Flashcard Converter

A collection of tools for creating language learning materials from SRT subtitles and CSV files. Generate Anki flashcards and PDF study materials with automatic vocabulary analysis and optional text-to-speech audio.

Features

• Extract vocabulary from SRT subtitle files using LLM analysis
• Import vocabulary from CSV files with flexible field mapping
• Generate Anki decks (.apkg) with:
  • Term, reading, meaning
  • Context sentences with translations
  • Optional text-to-speech audio
• Create PDF flashcards with ruby annotations
• Web interface for file uploads and processing
• Command-line interface for batch processing

Installation

This project uses the normal pyproject.toml configuration. You can use (uv)[https://astral.sh/uv/] to automatically handle environments and installation.

curl -LsSf https://astral.sh/uv/install.sh | sh

Usage

By default, the app uses Gemini Flash to handle translation and example sentence context. You'll need a Gemini API Key to run it locally. The quickest way to get a key is via AI Studio: https://aistudio.google.com/apikey. Set the key in your environment before running the app:

export GEMINI_API_KEY=...

Command Line Interface

You can generate flashcards in either Anki (apkg) or PDF format. When generating Anki packages, use the --audio flag to additionally generate TTS audio. TTS uses Google Cloud APIs, so you'll need to register for a Cloud Account setup the CLI and use gcloud auth application-default login if you want to generate audio.

The main entry points are flashcards_from_csv and flashcards_from_srt. When generating from a CSV, you can have the LLM infer a field mapping, or specify your own. By default the output file is written to out/[input].{pdf|apkg}.

uv run scripts/main.py flashcards_from_csv ./sample/n5.csv # inferred mapping
uv run scripts/main.py flashcards_from_csv ./sample/n5.csv --mapping=level=A,term=B

When generating an Anki package, you can specify the deck name and optionally output TTS audio for the term and translation:

uv run scripts/main.py flashcards_from_csv ./sample/n5.csv [--audio] [--deck-name=JLPT::N5]

You can exclude terms based on the term or translation by specifying a --filter argument:

uv run scripts/main.py flashcards_from_csv ./sample/n3.csv --filter=known.txt --filter=./sample/n4.csv --filter=./sample/n5.csv --output=pdf

You can change the language used for TTS audio using the --language flag:

uv run scripts/main.py flashcards_from_csv ./sample/n3.csv --language=japanese

SRTs operate the same way:

uv run scripts/main.py flashcards_from_srt ./saikik-ep11.srt --output=pdf

Web Interface

Start the web server:

uv run scripts/web.py

Navigate to http://localhost:8000 :

• Upload SRT files or CSV files
• Preview and customize field mappings
• Generate Anki decks or PDF flashcards
• Download processed files

License

Apache 2.0