Bespoke: Interactively Synthesizing Custom GUIs from Command-Line Applications By Demonstration
Despite their usability problems, command-line applications are frequently used by programmers, researchers, system administrators, and data scientists. To give these power users the well-known benefit of GUIs, we created Bespoke, an interactive system that synthesizes form-based GUIs by observing user demonstrations of command-line apps. To our knowledge, Bespoke is the first attempt to use demonstration-based techniques to synthesize GUIs from command-line invocations. To assess the versatility and limitations of Bespoke, we ran an exploratory study where we let participants use it to create a custom GUI in a domain that personally motivated them. They ended up creating a diverse set of GUIs for domains such as cloud computing management, prototyping machine learning, transcribing lecture videos, integrated circuit design, remote code deployment, and gaming server management. Participants reported that the main benefit of these bespoke GUIs is that they exposed only the most relevant subset of options required for their needs. In contrast, any official GUI made by the app's manufacturer would usually include far more panes, menus, and settings since they must accommodate a wider range of possible use cases.
Research Questions / Practical Problems
Question signals indexed for this paper.
- 71%
Where Will They Click Next? A Social Foraging Model for Collaborating Teams
CHI '26· Distributed Team Collaboration +2
- 71%
The Way We Notice, That’s What Really Matters: Instantiating UI Components with Distinguishing Variations
CHI '26· Human-LLM Collaboration +2
- 71%
Athena: Intermediate Representations for Iterative Scaffolded App Generation with an LLM
IUI '26· Human-LLM Collaboration +2
- 67%
Fork It: Supporting Stateful Alternatives in Computational Notebooks
CHI '21· Prototyping & User Testing +1
- 67%
Varv: Reprogrammable Interactive Software as a Declarative Data Structure
CHI '22· Prototyping & User Testing +1
- 67%
Learning to Denoise Raw Mobile UI Layouts for Improving Datasets at Scale
CHI '22· Prototyping & User Testing +1
- 67%
How To Draw Commands? An Elicitation Study for Sketching on Spreadsheets
CHI '25· Prototyping & User Testing +1
- 67%
Learning User Interface Semantics from Heterogeneous Networks with Multimodal and Positional Attributes
IUI '22· Prototyping & User Testing +1
- 67%
ODEN: Live Programming for Neural Network Architecture Editing
IUI '22· Prototyping & User Testing +1
- 67%
Relevance and Applicability of Hardware-independent Pointing Transfer Functions
UIST '21· Prototyping & User Testing +1
Based on Jaccard similarity of research subtopics & professions (≥60%)