Language Decoder

The LanguageDecoder translates a given text from the source language to a desired target language word by word ( decoding). For this purpose, the Google translator is used in the backend (Or alternatively an experimental LLM translator). To improve the translation, a dictionary can be created to automatically correct common translation errors. Afterwards, the decoded text can be exported to a PDF file.

The LanguageDecoder can also be used to create language learning lessons based on the Birkenbihl approach.

Usage

To use the LanguageDecoder you can either clone the repo with git:

git clone https://github.com/PumucklRandom/language-decoder.git

cd language-decoder

pip install -r requirements.txt

python ./__main__.py

or just download the latest desktop application from the releases.
For Windows use LanguageDecoder.zip, extract the files and execute LanguageDecoder.exe.
For Linux use LanguageDecoder-lx.zip, extract the files and execute LanguageDecoder.desktop.

With the build.py file you can create your own local desktop application.
(Install packages from requirements-dev.txt)

The LanguageDecoder requires a common modern browser such as Edge, Chrome or Firefox and a free port on localhost port 8080 (127.0.0.1:8080). If this port is occupied, the IP address and the port can be adjusted as desired in the config file: language-decoder/backend/config/config.yml; For the desktop application: LanguageDecoder/_internal/backend/config/config.yml

Known issues

Windows Security may block the desktop application. (Workaround: Run application as admin)
As an alternative on Windows, you may use the portable distribution LanguageDecoder-portable.zip and run LanguageDecoder.bat or LanguageDecoder.vbs

Credits

Nicegui for the grafical user interface
OpenRouter for the access of Large Language Models
Deep Translator for the language translation
FPDF2 for the pdf creation
TextStudio for the application icon