PosterGen
github.com
- Free plan
- Windows
- macOS
- Linux
- Web
- Self-hosted
What PosterGen is
PosterGen is free, MIT-licensed software that turns research paper PDFs into academic conference posters. Its multi-agent workflow uses a Parser Agent to extract and structure paper content, a Curator Agent to plan the narrative and visual placement, and Layout and Balancer agents to calculate spacing, arrange columns, and reduce overflow. Styling can derive colors from an affiliation logo, use professional fonts, and highlight keywords. The local web interface accepts a paper PDF and logos, lets users configure models and dimensions, then generates a PNG poster and an editable PowerPoint file. Windows, Linux, macOS, and self-hosted use are documented. Setup requires Python 3.11 and LibreOffice; the web interface also requires Node.js and configured API keys. OpenAI, Anthropic, and Zhipu keys are documented, with compatible base URLs including Google. Width-to-height ratios range from 1.4 to 2, and configuration supports layout, typography, colors, visual sizing, and content optimization.
Who it suits
It suits researchers and conference teams creating academic posters from paper PDFs. Users should be comfortable providing logos, configuring model access, and installing the stated dependencies.
Good to know, and what to watch
- Free and MIT-licensed
- Produces PNG and editable PowerPoint outputs
- Agent workflow structures content and layout
- Supports Windows, Linux, and macOS
- Custom dimensions and styling configuration
- Requires Python 3.11 and LibreOffice
- Web interface requires Node.js
- Configured API keys are required
- Requires a research PDF and logos
The verdict
PosterGen focuses on turning paper content into structured, editable conference posters through a configurable agent workflow. Its free license and cross-platform documentation are useful, but setup involves several dependencies, logos, and model credentials.
What the maker publishes
- Cost logging
- The output directory includes a JSON log for runtime and API cost. source
- Customization
- Users can customize layout, typography, color generation, visual asset sizing, and content optimization through the configuration file. source
- Dependencies
- The project acknowledges LangGraph, Marker, and python-pptx as underlying open-source projects. source
- Intended audience
- The README describes conference poster design and identifies the output as professional academic posters. source
- Intended users
- The README describes the system as generating academic posters and says it follows poster design practices for conferences. source
- Interface
- A local web interface lets users upload a PDF and logos, configure models and dimensions, then generate and download poster files. source
- Layout
- Its Layout Agent calculates positioning and spacing, while a Balancer Sub-Agent optimizes column use and prevents overflow. source
- License
- The repository identifies its license as MIT. source
- License
- The repository provides the software under the MIT License, which grants permission to use, copy, modify, distribute, sublicense and sell copies subject to its conditions. source
- License warranty
- The MIT license states that the software is provided as is without warranty. source
- Model options
- The documented text and vision model options include GPT-4.1, GPT-4o, GPT-4.1 mini and Claude Sonnet 4. source
- Model providers
- The README shows API key configuration for OpenAI, Anthropic and Zhipu and configurable base URLs for OpenAI, Anthropic, Google and Zhipu compatible services. source
- Models and APIs
- The README lists OpenAI, Anthropic, and Zhipu API keys and documents configurable API base URLs, including a Google base URL. source
- Operating systems
- The README lists Windows, Linux and macOS as supported operating systems and Python 3.11 as a system requirement. source
- Output
- Generated files include a PNG poster and an editable PowerPoint file. source
- Outputs
- The system produces PNG poster images and editable PowerPoint files. source
- Paper parsing
- Its Parser Agent extracts and structures content from a paper PDF. source
- Poster dimensions
- The documented width-to-height aspect ratio range is 1.4 to 2. source
- Purpose
- PosterGen generates academic posters from research papers using a multi-agent large language model framework guided by design principles. source
- Required inputs
- The setup instructions require a research paper PDF, an affiliation logo for color extraction and a conference logo for the poster. source
- Requirements
- The README specifies Python 3.11 and requires LibreOffice; its web interface also requires Node.js and configured API keys. source
- Storyboarding
- Its Curator Agent plans content organization and visual placement. source
- Styling
- Color generation draws on an affiliation logo, and typography features include professional font choices and keyword highlighting. source
- Workflow
- Its workflow uses parser, curator, layout and styling agents to extract paper content, plan a narrative storyboard, arrange a three-column layout and apply colors and typography. source
Checked against the maker's pages on 4 Oct 2026.