PaperBanana: AI Academic Illustration Generator

PaperBanana is an AI-driven tool designed to automate the creation of academic illustrations, transforming textual descriptions into publication-ready diagrams and charts. Targeting researchers and academics, it simplifies the visualization process by using a multi-agent architecture. The platform supports various figure types, including methodology diagrams, statistical plots, educational infographics, and system architectures. Key features include aesthetically refined visuals, accuracy in statistical representations via code generation, and the ability to enhance existing sketches. PaperBanana aims to reduce researchers' workload, improve the visual quality of publications, and make data presentation more precise and efficient.

📊 Accurate Statistical Plots
Generates Python Matplotlib code from data, ensuring bars, points, axes, and scales are mathematically precise, eliminating numerical hallucination and maintaining data-driven figure integrity. Useful if you need to represent experimental results in your paper.
🎨 Aesthetic Enhancement
Transforms rough sketches into professional illustrations that meet top-tier publication standards. By auto-summarizing aesthetic guidelines, systematically elevates color schemes, typography, spacing and visual quality for top venues such as NeurIPS, ICML, and ICLR
🤖 Multi-Agent Collaboration
Built on a five-agent architecture, PaperBanana delivers scientific figures that are faithful, precise, and aesthetically polished. This advanced framework enables the transformation of paper text into academic illustrations through sophisticated collaboration.