What does professional creative work actually look like, step by step? This sample dataset holds 294 annotated steps across 14 computer-use trajectories, recorded as working designers built real advertising and fashion-editorial assets in Adobe Photoshop and the browser.
We gave five designers from the Contra network a single creative brief each, a handheld-vacuum product launch, and recorded them carrying it through to multiple production-ready deliverables: an Instagram post, a vertical Story, a subway ad, a full-page print ad. The designers narrated their reasoning aloud as they worked, so every step's thought is transcribed from the expert's own voice.
Each step pairs a screenshot with a first-person thought, a structured action, and executable grounding: an Adobe Photoshop MCP tool call, a keyboard shortcut, a menu path, or a coordinate click, with a preferred execution path marking the most deterministic option. The format follows the AgentNet trajectory schema, extended with a Photoshop action taxonomy and multi-path execution. This structure makes the data directly usable in several training contexts:
- Computer-use agent SFT: long-horizon, multi-application trajectories with screenshots, actions, and coordinates in a format interoperable with existing GUI-agent tooling.
- Reasoning mid-training: pair screenshots with narration-grounded chain-of-thought to teach models how experts actually plan, act, and correct course mid-task.
- Tool-use and function calling: every step carries a structured tool_call, so agents can learn to reach the same goal through an API instead of pixel clicks.
- Benchmarking and evals: a human expert baseline for long-horizon creative workflows, executable end-to-end through the Photoshop MCP.
The dataset was recorded in 2026 on macOS and Windows. It is released under the CC BY 4.0 license on Hugging Face.