PaperBanana

Despite rapid advances in autonomous AI scientists, generating publication-ready illustrations remains a labor-intensive bottleneck. PaperBanana is an agentic framework that orchestrates specialized agents—Retriever, Planner, Stylist, Visualizer, and Critic—to transform raw scientific content into publication-quality diagrams and plots.

• Reference-Driven Generation

PaperBanana retrieves relevant reference examples to guide style and content, ensuring your diagrams match academic standards.

• Multi-Agent Collaboration

Five specialized agents work together: Retriever, Planner, Stylist, Visualizer, and Critic—each handling a critical step in the illustration pipeline.

• Iterative Self-Critique

The Critic agent inspects generated images against source content, providing feedback for automatic refinement until publication-ready.