v0.4.4

Which file do I download?

file	who it is for	what it does
transcript-normalizer-0.4.4-<system>-lite	anyone who just wants to use it	fetches the platform's captions and fixes domain terms; opens a menu when double-clicked
transcript-normalizer-0.4.4-<system>-full	anyone who also has videos without captions, or audio files	everything lite does, and also transcribes audio on your computer (the first time, it downloads a 1.5 GB speech model)
transcript_normalizer-0.4.4-py3-none-any.whl, .tar.gz	Python developers	the package, for pip install

<system> is windows, macos (Apple silicon) or linux. The executables are not signed, so Windows and macOS will warn the first time; the SHA-256 of every file is at the end of this page.

Qual arquivo baixar?

arquivo	para quem	o que faz
transcript-normalizer-0.4.4-<sistema>-lite	quem só quer usar	pega a legenda da plataforma e corrige os termos da área; abre um menu com dois cliques
transcript-normalizer-0.4.4-<sistema>-full	quem também tem vídeos sem legenda, ou arquivos de áudio	tudo o que o lite faz, e ainda transcreve o áudio no seu computador (na primeira vez, baixa um modelo de fala de 1,5 GB)
transcript_normalizer-0.4.4-py3-none-any.whl, .tar.gz	quem programa em Python	o pacote, para pip install

<sistema> é windows, macos (Apple Silicon) ou linux. Os executáveis não são assinados, então o Windows e o macOS avisam na primeira vez; o SHA-256 de cada arquivo está no fim desta página.

---

• Fix for the full executables: transcribing failed with "open() got an unexpected keyword argument 'metadata_errors'". The build had paired faster-whisper with a newer av it does not support. Executables are now built from the lock, the two have upper bounds, and every full executable is checked to decode audio before release (D-058).
• ? in the menu explains each option with its help entry and command, instead of repeating the menu.
• When the caption cannot be had, the menu says why in plain words (HTTP 429, or no caption) before asking about transcription, and does not show the video's details twice.
• The retry countdown stays on screen ("retrying in 4 s"), so the waits between attempts are visible.
• No Hugging Face warnings. The first model download says so in one line ("downloading the speech model (about 1.5 GB, first time only)…"), with a progress bar.

SHA-256

238fd1450e5925d7d978543e56d5c148ac5ec82b768973db8d8f7aabd2092bb1  transcript-normalizer-0.4.4-linux-full.tar.gz
091e47018b115abb50b22322ebbe61a81b8597c528e6908ea96f004d9e729ba0  transcript-normalizer-0.4.4-linux-lite.tar.gz
6a41daee743763a019f1ab7c66884b3b094a1cb24a0c0a2a4bc2c8a4e2f4f17b  transcript-normalizer-0.4.4-macos-full.zip
1129657ed2398dab35c4e78fd366b786d28f0544a2a3087497a1a1e93faa62fb  transcript-normalizer-0.4.4-macos-lite.zip
d7cd7ec72c9579b584312a54ed6e6af6dd152ed9bea54dd6e0ac84cbfc6c54f4  transcript-normalizer-0.4.4-windows-full.exe
952ebe8559a48d1a9a42c292bd158f6a3a8b5691aaceac24f174824146c3a140  transcript-normalizer-0.4.4-windows-lite.exe
3a41daf5e9ab8e006f1c68f14398598b4579a2f67de2a4bd0cd6fcadbb732600  transcript_normalizer-0.4.4-py3-none-any.whl
67f4f7cb77fdfe65e8d31649e0d9e49ea8a906943887c8251bbd368b9a808d63  transcript_normalizer-0.4.4.tar.gz