BigHugger
MCP MCP server · io.github.llmsresearch

io.github.llmsresearch/paperbanana

Generate academic diagrams and statistical plots from text using multi-agent AI.

host stars
2,372
30-day movement
+103/day
Related entries
2
Connections
0
pypilocalstdio

PaperBanana is an MCP server (stdio transport) that generates academic diagrams and statistical plots from text using a multi-agent AI pipeline of up to 7 specialized agents. It exposes CLI commands for methodology diagrams, matplotlib-based plots, batch generation, venue style packs, multi-panel composition, and quality evaluation.

Reach for it when you need publication-style figures and diagrams generated from text descriptions or paper sources without manual drawing.

Use it to

  • Generate methodology diagrams from text
  • Render statistical plots from CSV/JSON data via generated matplotlib code
  • Batch-generate figures from a YAML/JSON manifest
  • Compose multi-panel figures with (a)/(b)/(c) labels
  • Evaluate generated diagrams against human references

For Researchers and authors preparing academic paper figures

Runs
installed — on your machine
Transport
stdio
Packaged as
pypi:paperbanana
Install
uvx paperbanana
Version
0.1.1
Last release
2026-02-05
topicsacademic-figuresdiagramsplotsmulti-agentmatplotlibmcp-server