🚀 Make a video from your script

CS Board · Whiteboard Video Studio

Turn your ideas into a video that speaks. Add a reference voice and a Chinese script, then choose a visual template or provide character and style references. CS Board connects voice cloning, scene planning, illustration, hand-drawn animation, subtitles, and audio/video rendering into an MP4.

It runs as a local AI video workspace. Project files, API keys, job history, and rendered videos stay on your machine by default. A team on the same local network can share the production queue.

Mode	What you provide	Best for
Standard	Reference audio, Chinese script, visual template	Explainers, narrated stories, course promos.
Custom reference	Add a style image and character references	Consistent characters, branded series.
Dynamic infographic	Script and narration timing	Opinions, business analysis, teaching.

See the output

This animation is an example included in the repository. Browse the example assets.

The Chinese README also shows real screenshots of the voice library, style library, job history, task details, IndexTTS settings, and image generator. The 33 voices shown there were imported on the author's machine; they are not bundled with a fresh installation.

Install

Requirements

Component	Requirement
Operating system	Windows 10/11, or macOS 15+ for the current Remotion video renderer.
Runtime	Python 3.11+, Node.js 22.13+, FFmpeg and FFprobe on PATH.
Voice	A reachable IndexTTS 2.5 Gradio or FastAPI service.
Models	An OpenLux API key with access to text and image models.

On macOS, Homebrew can install the local runtimes:

brew install python@3.13 node ffmpeg

Clone this repository, then install dependencies from its root.

macOS:

git clone https://github.com/ChenShuo2004/cs-board.git
cd cs-board
python3.13 scripts/prepare_env.py
.venv/bin/python -m pip install -r webapp/requirements.txt
(cd web && npm ci)
./start-webapp.sh

Windows PowerShell:

git clone https://github.com/ChenShuo2004/cs-board.git
cd cs-board
python scripts/prepare_env.py
.\.venv\Scripts\python.exe -m pip install -r webapp\requirements.txt
Push-Location web
npm ci
Pop-Location
.\start-webapp.ps1

Open http://127.0.0.1:13000/. The web version and the macOS app use the same local ports (13000 and 18765), so quit CSBoard.app before starting the web version.

In API Settings, enter and test your OpenLux key, text and image models, and IndexTTS endpoint. Then upload a clean 10–30 second single-speaker voice sample and paste at least 10 Chinese characters. See the Chinese setup guide for optional image credentials, output paths, and TTS parameters.

What is included

Feature	Details
Voice library	Store, preview, search, rename, and reuse your own local reference voices.
Visual style library	12 built-in templates plus editable custom styles and preview images.
Character references	One style reference and up to five characters, with one to three images per character.
Rendering controls	16:9, 9:16, or 1:1; subtitle control and four drawing-density levels.
Job history	Search, inspect, reuse, cancel, and delete jobs; intermediate checkpoints support recovery.
Image generator	Make and download a PNG outside the video workflow at /image-generator.

For the full feature list and all 12 visual previews, see the Chinese README.

Repository and local data

assets/          Drawing and visual assets
docs/            Screenshots and workflow documentation
examples/        Whiteboard animation examples
scripts/         Environment setup and rendering utilities
tests/           Queue, recovery, and timing tests
video_renderer/  Remotion renderer
web/             React frontend
webapp/          FastAPI backend

Local configuration, reference voices, job files, and rendered videos are stored under .webapp/, which Git ignores. Keep API keys and private source media out of issues and commits; see SECURITY.md.

Star history

License

MIT · Chinese documentation · Follow the author on X