# llmsresearch/paperbanana

Open source implementation and extension of Google Research’s PaperBanana for automated academic figures, diagrams, and research visuals, expanded to new domains like slide generation.

Repository: https://github.com/llmsresearch/paperbanana
Canonical: https://ross.abutalabs.com/products/llmsresearch-paperbanana
Language: Python
License: MIT
License Family: permissive
Topics: academic-research, arxiv, gemini, llm, llms, mcp, mcp-server, multiagent, neurips, scientific-visualization, academic-diagrams, paperbanana, agentic-ai, diagram-generation, google-gemini, research-automation, research-tools, text-to-image, vlm, python-ai-research-tools
Last push: 2026-08-17T18:00:05+00:00

## Health v2 (maintenance only)
Score: 78/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 88, longevity 15
- inputs: {"age_days": 210, "days_push": 16, "days_rel": 83, "gap_med": 0, "n_releases_24m": 6}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2281, forks 333 (observed 2026-08-28T04:06:33.799521+00:00)

## What it is
An open-source, community implementation of Google Research's PaperBanana paper, providing an agentic Python framework that generates publication-quality academic diagrams, figures, and statistical plots from text descriptions using LLMs/VLMs like Gemini and OpenAI models. It also extends to new domains such as slide generation and can run as an MCP server.

## Use cases
- generate academic paper figures from text descriptions
- create publication-quality diagrams for a research paper
- make statistical plots for a NeurIPS submission
- generate slide decks and visuals from research notes
- render architecture diagrams for an arxiv paper
- automate scientific illustration with an LLM agent
- generate research visuals via an MCP server

## When to choose
- you need automated, publication-quality academic figures or diagrams from text
- you want an agentic, extensible pipeline for research visuals or slide generation
- you want to integrate diagram generation into LLM tooling via MCP

## When to avoid
- you need a general-purpose graphic design or photo editing tool
- you require exact fidelity to Google's internal original system, since this is an unofficial reimplementation
- you need offline generation without access to LLM APIs like Gemini or OpenAI

## Facets
- artifact type: library
- maturity: active
- function: agent-framework, image-processing, data-visualization, llm-inference, mcp, cli
- domain: artificial-intelligence, large-language-models, data-visualization, education, developer-tools
- platform: python, cli, cross-platform
- tags: academic-diagrams, diagram-generation, scientific-visualization, text-to-image, gemini, multiagent, vlm, research-automation, slide-generation, mcp-server, ai-agents

## Member repositories
- llmsresearch/paperbanana (main) score 78

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:33.799521+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:41:21.836838+00:00, confidence not recorded.
  - readme: https://github.com/llmsresearch/paperbanana (fetched 2026-08-28T04:06:33.799521+00:00, sha 741d562958f5)
  - registry_pypi: https://pypi.org/pypi/paperbanana/json (fetched 2026-08-29T10:21:34.539133+00:00, sha ca5eac40d633)
- Data as of 2026-08-30T08:39:29.467469+00:00.