screencastgen

pyproject.toml

Package metadata, dependencies, optional extras, and entry points.

Source: pyproject.toml

Package

Field Value
Name screencastgen
Version 2.0.0
Requires Python >= 3.9

Entry Points

Command Target
screencastgen screencastgen.cli:main → CLI
screencastgen-server screencastgen.inference_server:main → Inference Server

Optional Dependencies

Extra Packages What It Enables
qwen qwen-tts, torch, soundfile Qwen Backend
pydub pydub Concatenator (preferred)
client pydub, moviepy, Pillow, pymupdf Document/media processing on a remote-GPU client without local ML models
dev pytest, reportlab Unit tests and PDF test fixtures
docs mkdocs, mkdocs-material Local documentation preview and strict site builds
epub pydub, whisperx, torch, torchaudio, torchcodec EPUB Builder features
highlight pydub, whisperx, moviepy, Pillow, torch, torchaudio, torchcodec, pymupdf Highlight Pipeline including PDF page-image rendering via Page Renderer
lipsync pydub, whisperx, moviepy, Pillow, torch, torchaudio, torchcodec, pymupdf Lipsync Pipeline + reader bundle / MP4 assembly. Lip-sync providers such as LatentSync are installed separately.
server FastAPI, uvicorn, python-multipart, qwen, highlight, lipsync Inference Server
web FastAPI, uvicorn, SQLAlchemy, asyncpg, psycopg2, Alembic, Celery, Redis, sse-starlette, python-multipart, pydantic-settings Web backend/API/worker
gcs google-cloud-storage Storage Service GCS backend
s3 boto3 Storage Service S3 backend
all client + highlight + lipsync + qwen + server + web Full local Python stack

Core Dependency

Package Version Used By
PyPDF2 >= 3.0 Extractor

External Tools Not Declared As Extras

Tool Used By Note
manimgl executable Visualization Pipeline / ManimGL Renderer Must be installed separately and available on PATH
LatentSync sidecar env LatentSync Provider Installed by the local-GPU Setup Script, directly via scripts/install_latentsync.sh, or configured with LATENTSYNC_ROOT / LATENTSYNC_PYTHON

See Also

This site is open source. Improve this page.