#4343rd of 56 in Image and videoFor image, video and design, image and design, science, health and research, research and papers
paperbanana
Generates academic paper diagrams and plots from method text via a multi-agent pipeline
- Stars
- 2.5k
- License
- MIT
- Last commit
- Sep 2026
- Last release
- Jun 2026
- Language
- Python
1,314 on 21 Apr 20262,476 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Unofficial implementation of the PaperBanana paper. It turns methodology text or PDFs plus a caption into publication-style diagrams using a retriever, planner, stylist, visualizer and critic loop. Statistical plots are rendered from CSV/JSON via VLM-generated matplotlib code. It ships a CLI, Python API, MCP server, Claude Code skills and a local Gradio UI.
Who it is for: Researchers who need paper figures and plots generated from method text
Strengths
- Supports OpenAI, Azure, Gemini, Atlas Cloud and OpenRouter providers, selectable per run
- Batch manifests (YAML/JSON) with composite multi-panel figures and Markdown/HTML reports
- Accepts PDF input with page selection and reference sketch images to guide the planner
- Plots need only a VLM, no image-generation provider; venue style packs are customizable
Weaknesses
- Requires a cloud API key; no local model path is described
- README states it may differ from the original paper's system
- Gemini image model is listed at $0.134 per 1K image, so runs cost money
- Reference retrieval draws on a curated set of only 13 example diagrams
What it needs
- no GPU
- Docker
- Needs OpenAI, Azure OpenAI, Google Gemini, Atlas Cloud or OpenRouter API key, PyMuPDF (optional, for PDF input), Gradio (optional, for Studio)
- Models: gpt-5.2, gpt-image-1.5, gemini-2.5-flash, gemini-3-pro-image-preview, DeepSeek-V3-0324 (Atlas Cloud)
- port 7860
Quick answers
Is paperbanana open source?
Yes. Its code is on GitHub under the MIT license.
Can you run paperbanana with Docker?
Yes. Its repository has a Docker setup: a Dockerfile, a Compose file or a published image.
Does paperbanana need a GPU?
No, it runs without one.
Also in Image and video
See all 56| Rank | Project | Score |
|---|---|---|
| 1 | frigateLocal NVR with realtime object detection for IP cameras | 92 out of 100 |
| 2 | MoneyPrinterTurboGenerates short videos from a topic, with script, footage, voiceover and subtitles | 88 out of 100 |
| 3 | toonflow-appInfinite-canvas tool for AI short drama and comic video production | 86 out of 100 |
| #4 | ComfyUINode-graph editor for running diffusion image, video, audio and 3D workflows | 85 out of 100 |
| #5 | InvokeAIWeb UI for diffusion image generation with inpainting canvas and node workflows | 85 out of 100 |