TemplateCraft boosts one-click video and image content creation success
TemplateCraft: Agentic Visual Template Generation
MultimediaComputer Vision and Pattern Recognition
Summary
Creating reusable visual templates for videos and images is usually hard and time-consuming because it requires lots of manual work. The authors designed TemplateCraft, a system with multiple smart parts working together to turn simple instructions into ready-to-use templates that add effects to uploaded images or videos. TemplateCraft learns from trying and correcting mistakes without needing to retrain, which helps it make better templates over time. Tests show it improves how often images and videos are generated successfully and how well templates stick to their style.
What this means in practice
- •For content creation platforms: Generate reusable visual templates automatically to personalize images and videos with minimal manual work based on natural language input.$Commercial implications: Enables platforms to offer one-click personalized content creation tools that attract users with faster template generation and consistent styling.
- •For mobile app developers: Integrate agent-based template generation to automate effects and style workflows for image and video editing apps enhancing user experience.
Authors
Hongjie Yu, Zhiyuan Fan, Yuzhe Zhang, Jiangcun Du, Zhicheng Gao, Yuhong Zhang, Xiaokai Zhan, Zongshi Xie
Abstract
The growing popularity of short videos has driven demand for one-click content creation. Visual templates turn uploaded images into personalized content with preset effects, but reusable template generation still requires substantial manual effort in asset preparation and tool orchestration. We propose TemplateCraft, a multi-agent system that converts natural-language instructions into client-executable templates through planning, material generation, effect-workflow generation, and protocol compilation. Its Planner-Evaluator loop uses execution feedback for targeted rollback, while stage-level and long-term memory support revision without parameter updates. We evaluate TemplateCraft on TemplateBench, derived from 60 real-world templates. With the same Qwen3-VL backbone, TemplateCraft raises image/video generation success rates from 56.7%/30.0% to 66.7%/50.0% over Planner-only (best-of-three) and improves template adherence and style consistency. With additional evaluation and revision, it matches or exceeds a GPT-4o Planner-only baseline on selected metrics. Persistent assets further improve cross-input style consistency.