PaperBanana provides an agentic framework that automatically generates publication-quality academic diagrams and statistical plots from text descriptions, solving the time-consuming task of manual illustration creation for researchers. The server implements a two-phase multi-agent pipeline with iterative refinement capabilities and supports multiple vision language model and image generation providers including OpenAI, Azure OpenAI, and Google Gemini. It offers tools for input optimization and diagram generation with customizable parameters to produce figures suitable for academic papers and publications.
claude mcp add --transport stdio llmsresearch-paperbanana uvx paperbanana