Bibby's Paper Banana feature applies the multi-agent approach from Zhu et al. (2026) to turn your scientific content into publication-ready methodology diagrams and statistical plots — inside your Bibby workspace.
The overview below summarizes the original paper. To try Bibby's implementation, use Create Illustration.
Original paper authors
Abstract reproduced from Zhu et al. (arXiv:2601.23265). Bibby's tool is an independent implementation inspired by this work.
Despite rapid advances in autonomous AI scientists powered by language models, generating publication-ready illustrations remains a labor-intensive bottleneck in the research workflow. To lift this burden, we introduce PaperBanana, an agentic framework for automated generation of publication-ready academic illustrations.
Powered by state-of-the-art VLMs and image generation models, PaperBanana orchestrates specialized agents to retrieve references, plan content and style, render images, and iteratively refine via self-critique. To rigorously evaluate our framework, we introduce PaperBananaBench, comprising 292 test cases for methodology diagrams curated from NeurIPS 2025 publications.
Architecture from the PaperBanana paper — as implemented in Bibby's Paper Banana tool.
Capabilities reported in Zhu et al.; Bibby's implementation may differ in scope and availability.
Neural networks, flowcharts, multi-agent pipelines, and complex system architectures — all rendered to publication standards.
Accurate data visualization via Matplotlib code generation. Bar charts, ablation studies, and accuracy comparisons grounded in your data.
Transform rough hand-drawn sketches into clean, harmonious academic figures with consistent fonts and styling.
Upload existing diagrams to upgrade fonts, colors, and spacing without altering underlying content or structure.
Retrieves relevant papers to align style with academic conventions — your figures will match the venue aesthetic.
Self-critique loop ensures publication-quality output. The Critic agent reviews and forces regeneration until quality passes.
Benchmark results reported in the original paper (292 test cases from NeurIPS 2025). Figures are from Zhu et al., not Bibby-run evaluations.
The benchmark covers diverse research domains and illustration styles, representing the breadth of modern AI research publications.
Describe your methodology or paste your data. Bibby's Paper Banana applies the paper's agent workflow inside your project.
Input your methodology, data, or sketch. PaperBanana accepts text, PDFs, or images.
The Retriever agent scans reference databases to find style-aligned academic examples.
Planner and Stylist agents create a detailed visual plan with academic-grade aesthetics.
Visualizer renders the figure; Critic reviews and iterates until quality is publication-ready.
Cite Zhu et al. when referring to the research framework. Bibby's web tool is a separate product implementation.
@article{zhu2026paperbanana,
title={PaperBanana: Automating Academic
Illustration for AI Scientists},
author={Zhu, Dawei and Meng, Rui and
Song, Yale and Wei, Xiyu and
Li, Sujian and Pfister, Tomas and
Yoon, Jinsung},
journal={arXiv preprint arXiv:2601.23265},
year={2026}
}Original research affiliations (not Bibby)
Need help with academic illustrations or want a custom feature?
Book a Free Call