LaTCoder: Converting Webpage Design to Code with Layout-as-Thought

TL;DR

LaTCoder is a novel method that improves webpage design to code conversion by enhancing layout preservation through a layout-as-thought approach, utilizing chain-of-thought prompting and strategic assembly, leading to significant accuracy improvements.

Contribution

We introduce LaTCoder, a new approach that combines layout segmentation, chain-of-thought prompting, and assembly strategies to better preserve webpage layouts during code generation.

Findings

TreeBLEU scores increased by 66.67% with DeepSeek-VL2.

MAE decreased by 38% on the new CC-HARD benchmark.

Over 60% human preference for LaTCoder-generated webpages.

Abstract

Converting webpage designs into code (design-to-code) plays a vital role in User Interface (UI) development for front-end developers, bridging the gap between visual design and functional implementation. While recent Multimodal Large Language Models (MLLMs) have shown significant potential in design-to-code tasks, they often fail to accurately preserve the layout during code generation. To this end, we draw inspiration from the Chain-of-Thought (CoT) reasoning in human cognition and propose LaTCoder, a novel approach that enhances layout preservation in webpage design during code generation with Layout-as-Thought (LaT). Specifically, we first introduce a simple yet efficient algorithm to divide the webpage design into image blocks. Next, we prompt MLLMs using a CoTbased approach to generate code for each block. Finally, we apply two assembly strategies-absolute positioning and an MLLM-based method-followed by dynamic selection to determine the optimal output. We evaluate the effectiveness of LaTCoder using multiple backbone MLLMs (i.e., DeepSeek-VL2, Gemini, and GPT-4o) on both a public benchmark and a newly introduced, more challenging benchmark (CC-HARD) that features complex layouts. The experimental results on automatic metrics demonstrate significant improvements. Specifically, TreeBLEU scores increased by 66.67% and MAE decreased by 38% when using DeepSeek-VL2, compared to direct prompting. Moreover, the human preference evaluation results indicate that annotators favor the webpages generated by LaTCoder in over 60% of cases, providing strong evidence of the effectiveness of our method.