Pixelle Video
github.com/aidc-ai/pixelle-video- Category
- AI Tools
- Rank
- No. 765Tools index
Previous survey · No. 751 ·
- Type
- APP
- Builder
- aidc-ai
- GitHub
- 28.1k stars
- Latest release
- v0.1.15
- Date
About
A fully automated AI engine for generating short-form videos end to end.
What it does
Pixelle Video turns a topic or finished script into a narrated video through configurable stages. It can write narration, plan scenes, generate images or clips, synthesize speech, add music, and assemble the result. A browser interface exposes model, voice, dimensions, style, template, and workflow choices.
Why it's ranked here
Its strongest case is breadth with replaceable parts. Users can mix local ComfyUI workflows, RunningHub services, direct media APIs, several speech systems, and OpenAI-compatible language models. That flexibility is substantial, but setup complexity and reliance on external services make the experience less predictable than the simple topic-first pitch suggests.
What's good
It supports both generated and supplied scripts, so automation does not force users to surrender editorial control. Text-only templates can avoid ComfyUI, while richer pipelines can use local or cloud generation. Custom HTML templates and exported ComfyUI workflows provide practical extension points. Voice previews, style previews, progress reporting, and completed-video metadata help shorten iteration cycles.
Tradeoffs
Source installation requires Python 3.11 or newer, a package manager, and FFmpeg. Some browser-dependent features also require Chrome. Rich media workflows may need local ComfyUI models, cloud credentials, provider balances, or capable hardware. The default speech service can fail under network instability. Documentation also conflicts on whether the license is Apache 2.0 or MIT.
How to use it well
Use it for repeatable narrated shorts, explainers, announcements, and template-driven content where scripts, visuals, speech, and music follow a stable pattern. Start with supplied copy and a text-only template, then add generated media after validating timing and narration. It does not replace detailed timeline editing or guarantee reliable generation across third-party services.
Technical notes+
pyproject.toml defines a Python 3.11 package built with Hatchling and includes Streamlit, FastAPI, Uvicorn, FastMCP, MoviePy, Playwright, ComfyKit, and provider SDK dependencies. README.md describes a Streamlit browser workflow, direct media-provider configuration, ComfyUI and RunningHub integration, and FFmpeg-based source setup. docs/en/user-guide/api.md documents an asynchronous Python library interface through PixelleVideoCore. docs/FAQ.md describes workflow prefix conventions and network-related speech failures, but incorrectly calls the license MIT; README.md and pyproject.toml specify Apache 2.0.
Observed
- License
- Apache-2.0 in project metadata and README; FAQ documentation contains conflicting MIT text.
- Primary language
- Python 3.11 or newer.
- Packaging
- Hatchling build backend with uv or traditional pip and virtual environments for source installation.
- Interfaces
- Streamlit web interface and asynchronous Python library API.
- Platform support
- Prebuilt Windows package; documented source installation for macOS and Linux.
- Media dependencies
- FFmpeg is required for source installation, and some browser-dependent features require Chrome.
Read from README.md, pyproject.toml, docs/FAQ.md, docs/FAQ_CN.md, docs/en/faq.md, docs/zh/faq.md, docs/en/index.md, docs/zh/index.md, docs/gallery/index.md, docs/en/troubleshooting.md, docs/zh/troubleshooting.md, docs/en/gallery/index.md, docs/zh/gallery/index.md, docs/en/user-guide/api.md, docs/zh/user-guide/api.md.
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.