Metadata-Version: 2.5
Name: legendaria
Version: 0.1.0.dev2
Summary: Legendária transcription engine: local Whisper transcription with word timings (CLI: flosub).
Project-URL: Homepage, https://github.com/flou-ainan/Legendaria
Project-URL: Issues, https://github.com/flou-ainan/Legendaria/issues
Author: flou-ainan
License-Expression: MIT
License-File: LICENSE
Keywords: captions,srt,subtitles,transcription,vtt,whisper
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: Environment :: Console
Classifier: Intended Audience :: End Users/Desktop
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Multimedia :: Sound/Audio :: Speech
Classifier: Topic :: Multimedia :: Video
Requires-Python: >=3.10
Requires-Dist: faster-whisper<2,>=1.2
Requires-Dist: platformdirs>=4
Provides-Extra: cuda
Requires-Dist: nvidia-cublas-cu12>=12.1; (sys_platform == 'linux' or sys_platform == 'win32') and extra == 'cuda'
Requires-Dist: nvidia-cudnn-cu12<10,>=9; (sys_platform == 'linux' or sys_platform == 'win32') and extra == 'cuda'
Provides-Extra: test
Requires-Dist: jsonschema>=4; extra == 'test'
Requires-Dist: pytest>=8; extra == 'test'
Description-Content-Type: text/markdown

# Legendária engine (`flosub`)

The transcription engine of [Legendária](https://github.com/flou-ainan/Legendaria): it runs Whisper (faster-whisper) locally and writes subtitles with per-word timings. The Blender add-on uses it, and it also works on its own:

```bash
flosub transcribe video.mp4                  # writes video.json, video.srt, video.vtt
flosub transcribe talk.mp3 --language pt --model medium --format srt
flosub export video.json --format vtt
flosub models list
flosub models download medium
flosub config show
```

GPU acceleration works on NVIDIA cards (Linux and Windows) with the `cuda` extra; everything else runs on the CPU.

## Development

```bash
cd engine
uv sync --extra test             # add --extra cuda for NVIDIA GPUs
uv run flosub --help
uv run pytest
```

License: MIT.
