Figma2Code: Automating Multimodal Design to Code in the Wild

TL;DR

Figma2Code introduces a multimodal approach to convert Figma design files into UI code, leveraging rich metadata to improve automation in real-world development workflows.

Contribution

The paper presents a new multimodal dataset and benchmarks for design-to-code tasks using Figma files, advancing beyond image-only approaches.

Findings

Proprietary models excel in visual fidelity but struggle with layout responsiveness.

Models tend to directly map primitive visual attributes from metadata.

A curated dataset of 213 high-quality Figma cases was created for benchmarking.

Abstract

Front-end development constitutes a substantial portion of software engineering, yet converting design mockups into production-ready User Interface (UI) code remains tedious and costly. While recent work has explored automating this process with Multimodal Large Language Models (MLLMs), existing approaches typically rely solely on design images. As a result, they must infer complex UI details from images alone, often leading to degraded results. In real-world development workflows, however, design mockups are usually delivered as Figma files, a widely used tool for front-end design, that embed rich multimodal information (e.g., metadata and assets) essential for generating high-quality UI. To bridge this gap, we introduce Figma2Code, a new task that advances design-to-code into a multimodal setting and aims to automate design-to-code in the wild. Specifically, we collect paired design images and their corresponding metadata files from the Figma community. We then apply a series of processing operations, including rule-based filtering, human- and MLLM-based annotation and screening, and metadata refinement. This process yields 3,055 samples, from which designers curate a balanced dataset of 213 high-quality cases. Using this dataset, we benchmark ten state-of-the-art open-source and proprietary MLLMs. Our results show that while proprietary models achieve superior visual fidelity, they remain limited in layout responsiveness and code maintainability. Further experiments across modalities and ablation studies corroborate this limitation, partly due to models' tendency to directly map primitive visual attributes from Figma metadata.