What problem does it solve?
It eliminates the manual, error-prone work of converting experiment outputs into consistent, publication-ready figures and tables for a paper.
Core Features & Use Cases
- Data-driven figure generation: Creates line plots, bar charts, scatter plots, heatmaps, and box/violin plots from experiment data for typical ML paper workflows.
- Paper comparison artifacts: Produces comparison tables (LaTeX) for prior work and ablation/baseline results, plus multi-panel figures for multi-dataset and multi-metric layouts.
- Style consistency for publication: Applies a configurable publication style (DPI, fonts, palette, vector PDF) and generates LaTeX include snippets to speed up manuscript assembly.
- Preserves manual hero figures: Detects existing manually created figures and keeps them, generating only what is data-driven.
Quick Start
Generate paper figures from your figure plan by running the paper-figure skill with your figure plan file or data path as the single argument.