Talk to Your Slides: High-Efficiency Slide Editing via Language-Driven Structured Data Manipulation

TL;DR

This paper introduces Talk-to-Your-Slides, a language-driven slide editing system that manipulates structured data for efficient, precise, and style-preserving presentation editing, outperforming visual agents in speed and fidelity.

Contribution

It presents a novel hierarchical architecture for slide editing via language, focusing on structured data manipulation instead of visual perception, with a new benchmark dataset.

Findings

34% faster processing for text-centric tasks

34% better instruction fidelity

87% lower operational cost compared to GUI baselines

Abstract

Editing presentation slides is a frequent yet tedious task, ranging from creative layout design to repetitive text maintenance. While recent GUI-based agents powered by Multimodal LLMs (MLLMs) excel at tasks requiring visual perception, such as spatial layout adjustments, they often incur high computational costs and latency when handling structured, text-centric, or batch processing tasks. In this paper, we propose Talk-to-Your-Slides, a high-efficiency slide editing agent that operates via language-driven structured data manipulation rather than relying on the image modality. By leveraging the underlying object model instead of screen pixels, our approach ensures precise content modification while preserving style fidelity, addressing the limitations of OCR-based visual agents. Our system features a hierarchical architecture that effectively bridges high-level user instructions with low-level execution codes. Experiments demonstrate that for text-centric and formatting tasks, our method enables 34% faster processing, achieves 34% better instruction fidelity, and operates at an 87% lower cost compared to GUI-based baselines. Furthermore, we introduce TSBench, a human-verified benchmark dataset comprising 379 instructions, including a Hard subset designed to evaluate robustness against complex and visually dependent queries. Our code and benchmark are available at https://github.com/KyuDan1/Talk-to-Your-Slides.