Apps · PROJECT NOTES
VideoGUI
Instructional-video GUI benchmark data for creative software, separating high-level planning, subtask planning and individual mouse/keyboard actions.
What can you do with it?
Study reconstruction of visual workflows in tools such as Photoshop, Premiere, After Effects, DaVinci Resolve and web creation interfaces. Inspect released high-plan, mid-plan and action data as separate evaluation units.
What should you know before using it?
Data is public, but the repository explicitly leaves evaluation code and metric implementation unreleased. It does not provide a turnkey application VM or reset-and-score harness. The paper reports 86 full tasks and 463 subtasks; the reviewed Hugging Face release exposes 82 high-plan metadata files, 462 mid-plan rows and 2,572 action rows, which are different units and must not be summed. No explicit license was found in the reviewed repository or dataset cards. Replaying professional applications requires suitable installations and licenses.
What is verified here?
Runtime: Released recordings, screenshots and structured planning/action annotations; no released VideoGUI metric implementation was established. Windows/web tags denote source task interfaces, not validated host portability.
License note: No explicit code/data license was declared in the reviewed repository or the three linked Hugging Face dataset cards.
This entry was reviewed against original project material. We have not independently run the software or reproduced its results. AI-development claims are author-reported; check upstream for current versions, setup requirements, and licenses.
- Resource type
- Dataset + benchmark
- Provenance
- Development provenance not established
- Reviewed
- Platforms
- Windows, Web