Codex Whiteboard Video Skill

Generates hand-drawn whiteboard videos from images or scripts by connecting the Codex platform to a local rendering engine.

Generates hand-drawn whiteboard videos from images or scripts by connecting the Codex platform to a local rendering engine.

The gist

Codex Whiteboard Video Skill is an open-source adapter layer created by gnipbao. It enables the Codex platform to generate hand-drawn whiteboard style videos by calling a separate, locally-installed Python engine. This skill bridges the gap between Codex's command interface and the underlying whiteboard-video-engine, allowing users to convert images, SVGs, line art, or scripts into animated videos directly from their workflow.

What it does

  • Converts images, SVGs, line art, and scripts into hand-drawn whiteboard videos.
  • Integrates the whiteboard-video-engine into the Codex platform as a callable skill.
  • Customizes video output with parameters for duration, line art detail, and hand gesture style.
  • Supports multiple external, user-provided models for line art generation, such as Informative Drawings and Anime2Sketch.

How it works

A user installs the whiteboard-video-engine via pip, then clones this repository into their local Codex skills folder. From within Codex, they provide a prompt that references the skill and an input file. This triggers a Python wrapper script that calls the engine's command-line interface to render an MP4 video. The tool is open-source and requires separate installation and configuration of the core engine and its models.

Best for

Developers and technical users of the Codex platform who need to programmatically generate whiteboard explainer videos from static assets like images or SVGs.

Watch out for

This is not a standalone application but a lightweight wrapper. It requires separate installation of the whiteboard-video-engine and its associated machine learning models, which must be configured correctly to function.