Papers.Data.Code
Dataset

FigmaTrace

PatronusAI

Contains 3,469 image-text action trajectories from expert Figma sessions across 126 long-horizon design tasks. Each sequence pairs screen frames with Playwright-MCP actions and phase labels.

Hugging Face ↗Aug 19, 2026

  • multimodal
  • gui-agents
  • figma
  • trajectories
Taxonomy of skills characterizing the work's scope and structure
Taxonomy of skills characterizing the work's scope and structure

What it is

Expert Figma workflows captured as action-frame trajectories, turning 200+ hours of design work into trainable GUI-agent data. 3,469 action-frame trajectories from 200+ hours of expert Figma sessions across 126 long-horizon tasks, with 12 phase labels, 10-skill taxonomy, and 8 workflow categories.

Key results

  • 3,469 trajectories; 2,883 train and 586 eval
  • 126 tasks across 8 workflow categories
  • 200+ hours of screen recordings; CC-BY-4.0

Useful for supervised fine-tuning and evaluation of VLM-based GUI agents for Figma design workflows.

One useful ML email a week.

More from Aug 24 – Aug 30

The whole week →

Picked Aug 28

Picked Aug 27

  • Dataset
    open-gigaai/GigaBrain-0.7-SampleData

    Provides permissively licensed sample VLA data linked to large-scale multi-embodiment training for studying embodied foundation model pipelines.

    Hugging FaceAug 25, 2026
    →
  • Repo
    ApodexAI/FrontierAgent

    Unlike most agent runtimes, it ships built-in evaluation and shows parallel Agent Team beating ReAct 38.5 vs 34.4 on APEX.

    GitHubAug 22, 2026
    →

Picked Aug 25

  • Repo
    Leonxlnx/unlazy

    Unlike orchestration runtimes, it enforces file-based completion gates and stop blocking, reporting 1.6-3.9× more effort and 4-10 self-found defects fixed.

    GitHubAug 9, 2026
    →

Worth your time, not all of your time.

We track new ML work across papers, code, and data, then keep only the pieces worth a closer look. Selected, not collected.

Papers.Data.Code

The Weekly ML Signal

Selected, not collected.

One useful ML email a week: the papers, repos and datasets worth your attention.

Worth your time, not all of your time.