Code2Video: A Code-centric Paradigm for Educational Video Generation
Paper · https://arxiv.org/abs/2510.01174
While recent generative models advance pixel-space video synthesis, they remain limited in producing professional educational videos, which demand disciplinary knowledge, precise visual structures, and coherent transitions, limiting their applicability in educational scenarios. Intuitively, such requirements are better addressed through the manipulation of a renderable environment, which can be explicitly controlled via logical commands (e.g., code). In this work, we propose Code2Video, a code-centric agent framework for generating educational videos via executable Python code. The framework comprises three collaborative agents: (i) Planner, which structures lecture content into temporally coherent flows and prepares corresponding visual assets; (ii) Coder, which converts structured instructions into executable Python codes while incorporating scope-guided auto-fix to enhance efficiency; and (iii) Critic, which leverages vision-language models (VLM) with visual anchor prompts to refine spatial layout and ensure clarity. To support systematic evaluation, we build MMMC, a benchmark of professionally produced, discipline-specific educational videos. We evaluate MMMC across diverse dimensions, including VLM-as-a-Judge aesthetic scores, code efficiency, and particularly, TeachQuiz, a novel end-to-end metric that quantifies how well a VLM, after unlearning, can recover knowledge by watching the generated videos. Our results demonstrate the potential of Code2Video as a scalable, interpretable, and controllable approach, achieving 40% improvement over direct code generation and producing videos comparable to human-crafted tutorials. The code and datasets are available at https://github.com/showlab/Code2Video.
First seen by Oriel on .
AI summary
Code2Video is a code-centric multi-agent framework that generates controllable educational videos through executable Python, reporting 40% improvement over direct code generation and quality comparable to human-crafted tutorials.
- Planner structures lecture content into temporally coherent flows and prepares visual assets, while Coder translates its instructions into executable Python code.
- Coder uses scope-guided auto-fix for efficiency, and Critic applies VLMs with visual anchor prompts to improve spatial layout and clarity.
- The authors created MMMC, a benchmark of professionally produced, discipline-specific educational videos evaluated on aesthetics, code efficiency, and teaching effectiveness.
- TeachQuiz measures whether a VLM, after unlearning, can recover knowledge by watching generated videos.
- The framework targets educational content requiring disciplinary knowledge, precise visual structures, and coherent transitions that pixel-space video synthesis struggles to provide.
#educational video #multi-agent systems #code generation #vision-language models
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Hugging Face Papers: 36, latest observation ; 0 over observed period since
Comments
Comment count reported by the source.
- Hugging Face Papers: 4, latest observation ; 0 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hugging Face Papers | 2510.01174 | https://arxiv.org/abs/2510.01174 Read discussion |