Pixelle-Video is an open source AI video generator that turns one topic into a finished short video workflow. It is built for creators who need publishable clips fast, but still want control over models, prompts, and hosting choices. The project runs through a Streamlit interface and coordinates script writing, visual generation, voiceover, background music, and final rendering in one place.
It fits solo creators, growth teams, and educators who produce frequent short-form explainers or social content. Instead of stitching together five separate tools, you can keep one pipeline and swap components as needed. The repository includes both self-hosted and RunningHub workflow configurations, so teams can start locally and move to hosted GPU capacity when volume grows.
Key Features
- Generate a full video pipeline from a single theme input.
- Draft narration scripts automatically, then adjust tone before render.
- Produce scene visuals with image or video generation workflows.
- Add TTS voice tracks and background music in the same run.
- Export vertical or horizontal formats for different channels.
- Extend the flow with ComfyUI-based modular workflow files.
Self-Hosting
Pixelle-Video can run as a local Streamlit app with "uv run streamlit run web/app.py" after installing dependencies such as uv and ffmpeg. A Windows package is also available for teams that want a faster setup path with fewer manual steps.
License
Pixelle-Video is released under Apache-2.0. This license allows commercial use, modification, and redistribution with notice and license preservation requirements.
