PSDesigner: Automated Graphic Design with a Human-Like Creative Workflow
Authors: Xincheng Shuai, Song Tang, Yutong Huang, Henghui Ding, Dacheng Tao
Organizations: Institute of Big Data, College of Computer Science and Artificial Intelligence, Fudan University, China · Generative AI Lab, College of Computing and Data Science, Nanyang Technological University, Singapore
Graphic design is a creative and innovative process that plays a crucial role in applications such as e-commerce and advertising. However, developing an automated design system that can faithfully translate user intentions into editable design files remains an open challenge. Although recent studies have leveraged powerful text-to-image models and MLLMs to assist graphic design, they typically simplify professional workflows, resulting in limited flexibility and intuitiveness. To address these limitations, we propose PSDesigner, an automated graphic design system that emulates the creative workflow of human designers. Building upon multiple specialized components, PSDesigner collects theme-related assets based on user instructions, and autonomously infers and executes tool calls to manipulate design files, such as integrating new assets or refining inferior elements. To endow the system with strong tool-use capabilities, we construct a design dataset, CreativePSD, which contains a large amount of high-quality PSD design files annotated with operation traces across a wide range of design scenarios and artistic styles, enabling models to learn expert design procedures. Extensive experiments demonstrate that PSDesigner outperforms existing methods across diverse graphic design tasks, empowering non-specialists to conveniently create production-quality designs.
Figures & tables
Figure 2: The typical layer hierarchy in PSD (Adobe Photoshop Document) files, where the layers used to compose the same visual concept ( e.g . , “Left Panel”) are grouped together.
Figure 3: The three-stage construction pipeline of the proposed design dataset CreativePSD. We first collect high-quality PSD files from the internet and paid data, while grouping the layers based on their underlying visual concepts. Then, we parse the PSD files and extract essential information, such as raw assets, metadata, and intermediate renders. Finally, we use the extracted data to construct the training data for Xgen and Xedt modes of GraphicPlanner.
Figure 4: Evaluation of the model performance on translating user intentions to the final designs. Most of the compared methods can only generate non-editable raster images or output a few layers with simple attributes, hindering their professionalism and flexibility. Furthermore, many methods can not generate accurate texts, especially for complex characters, e.g . , Chinese.
Figure 5: Evaluation of model performance on the graphic design composition task, using Crello-v5 dataset [ 50 ] . Our method achieves coherent arrangements, achieving visually appealing outcomes.
Metric
Ours
OpenCOLE [ 64 ]
Bagel [ 74 ]
FLUX [ 14 ]
PosterCraft [ 75 ]
CanvaGPT
VLM
Human
VLM
Human
VLM
Human
VLM
Human
VLM
Human
VLM
Human
Qua.
7.62
8.09
5.12
3.95
5.85
6.12
8.18
7.68
7.95
8.35
8.52
8.28
Lay.
8.68
9.12
6.85
4.58
4.47
6.88
3.66
5.41
6.39
8.15
8.35
7.08
Rel.
7.78
7.17
5.25
3.48
6.58
5.24
6.92
8.15
8.42
7.10
4.72
6.88
Har.
8.02
7.72
6.68
5.92
7.25
6.08
6.82
5.02
8.05
7.65
7.21
7.88
Inn.
8.45
6.98
6.08
7.32
5.68
6.45
6.95
7.18
5.87
7.58
7.52
6.33
Table 2: VLM and human evaluation scores for graphic design. All scores are within the range of 1 to 10.
Method
Qua.
Lay.
Har.
Inn.
Ours
7.85
7.43
6.77
6.94
LaDeCo [ 5 ]
5.95
6.03
7.22
5.75
Ground Truth
8.13
9.18
8.90
7.12
Table 3: VLM evaluation scores for graphic design composition, using Crello-v5 [ 50 ] dataset.
Method
Qua.
Lay.
Har.
Inn.
Crello-v5
Ours
7.85
7.43
6.77
6.94
Ours w/o Xedt
6.05
5.88
5.90
6.75
Ours w/o M
6.25
6.10
6.18
6.02
Ours w/o RL
6.38
6.00
6.35
6.20
PSD
Ours
6.28
6.15
7.02
6.88
Ours w/o Xedt
5.32
5.15
6.22
6.05
Table 4: Ablation studies.
Figure 6: The results of our method in graphic design composition tasks with complex layer hierarchies.
We introduce GraphicDesignBench (GDB), the first comprehensive benchmark suite designed specifically to evaluate AI models on the full breadth of professional graphic design tasks. Unlike existing benchmarks that focus on natural-image understanding or generic text-to-image synthesis, GDB targets the unique challenges of professional design work: translating communicative intent into structured layouts, rendering typographically faithful text, manipulating layered compositions, producing valid vector graphics, and reasoning about animation. The suite comprises 50 tasks organized along five axes: layout, typography, infographics, template & design semantics and animation, each evaluated under both understanding and generation settings, and grounded in real-world design templates drawn from the LICA layered-composition dataset. We evaluate a set of frontier closed-source models using a standardized metric taxonomy covering spatial accuracy, perceptual quality, text fidelity, semantic alignment, and structural validity. Our results reveal that current models fall short on the core challenges of professional design: spatial reasoning over complex layouts, faithful vector code generation, fine-grained typographic perception, and temporal decomposition of animations remain largely unsolved. While high-level semantic understanding is within reach, the gap widens sharply as tasks demand precision, structure, and compositional awareness. GDB provides a rigorous, reproducible testbed for tracking progress toward AI systems that can function as capable design collaborators. The full evaluation framework is publicly available.
Professional designers work from client briefs that specify goals and constraints but often lack concrete design details. Translating these abstract requirements into visual designs poses a central challenge, yet existing tools address specific aspects or induce fixation through complete outputs. Through interviews with six professional designers, we identified how designers address this challenge: first structuring ambiguous requirements, then exploring individual elements, and finally recombining alternatives. We developed Brief2Design, supporting this workflow through requirement extraction and recommendation, element-level exploration for objects, backgrounds, text, typography, and composition, and flexible recombination of selected elements. A within-subjects study with twelve designers compared Brief2Design against a conversational baseline. The structured approach increased prompt diversity and received high ratings for requirement extraction and recommendation, but required longer generation time and achieved comparable image diversity. These findings reveal that structured workflows benefit requirement clarification at the cost of efficiency, informing design trade-offs for AI-assisted graphic design tools.
Designing visually diverse and high-quality designs remains a manual, time-consuming process, limiting scalability and personalization in creative workflows. We present a system for generating editable design variations using a decoder-only language model, the Creative Pre-trained Transformer (CPT), trained to predict visual style attributes in design templates. At the core of our approach is a new representation called Creative Markup Language (CML), a compact, machine-learning-friendly format that captures canvas-level structure, page layout, and element-level details (text, images, and vector graphics), including both content and style. We fine-tune CPT on a large corpus of design templates authored by professional designers, enabling it to learn meaningful, context-aware predictions for attributes such as color schemes and font choices. The model produces semantically structured and stylistically coherent outputs, preserving internal consistency across elements. Unlike generative image models, our system yields fully editable design documents rather than pixel-only images, allowing users to iterate and personalize within a design editor. In experiments, our approach generates contextual color and font variations for existing templates and shows promise in adjusting layouts while maintaining design principles.