Synthux Economy R9 W223: Simulated Desktop Computer-Use Trajectories

Grounded visual computer-use trajectories generated by the SynthUX system. Each record captures one worker's device session inside a simulated company, where goals drive interactions with applications like Terminal, VS Code, and Slack through low-level mouse and keyboard inputs. The dataset was created by jacob-valdez and was last updated on June 4, 2026.

Use Cases

• Train or benchmark AI agents for desktop task automation based on the described low-level mouse and keyboard input sequences.
• Study human-like computer interaction patterns based on the goal-driven, node-tree structure of the recorded sessions.
• Develop models for screen understanding and action prediction based on the recorded screen video and per-node frames.
• Simulate and analyze workplace productivity workflows based on interactions with applications like VS Code, Slack, and Sheets mentioned in the description.

Strengths

• Data includes multi-modal recordings of screen video and per-node frames for each session.
• Trajectories are grounded in a simulated company environment with a described goal-to-node-tree expansion structure.
• Interactions span a variety of common desktop applications mentioned, including Terminal, VS Code, Browser, and Slack.

Limitations

• Description metadata is limited; actual data quality, scale, and column definitions require manual inspection after download.
• Row count, file formats, and license information are unknown, which may limit suitability assessment.
• Data is synthetic, generated by the SynthUX simulator, and may not fully reflect real-world user behavior or complexity.

Provenance

Source

jacob-valdez on Hugging Face

Collection Method

Generated by the SynthUX system using a desktop simulator to record goal-driven user sessions.

Freshness

Last updated 2026-06-04 07:34:31; freshness should be verified.