Posters.science

A free, open-source platform for sharing, discovering, and citing scientific posters.

Website · Developer Docs · Report Bug · Request Feature

---

About

Posters.science is a platform for researchers to upload, share, and discover scientific conference posters. When a poster is uploaded, the platform automatically extracts structured metadata such as titles, authors, affiliations, sections, and figure captions. This makes posters findable, citable, and machine-readable.

The platform is built around FAIR principles (Findable, Accessible, Interoperable, Reusable) and integrates with Zenodo so that posters can be deposited with a DOI for long-term archival and citation.

Posters.science is developed by the FAIR Data Innovations Hub at the California Medical Innovations Institute (CalMI2).

Related Resources

Resource	Description
poster2json	Python package and CLI for poster metadata extraction (docs)
poster-json-schema	JSON schema for machine-actionable and FAIR poster metadata (DataCite 4.7)
poster-json-examples	Manually annotated ground-truth poster examples
posters-science-extraction-api	Extraction API service used by the platform
poster-sentry	Lightweight multimodal scientific poster classifier
poster-sentry-training	Training data and scripts for the poster-sentry classifier
posters-science-dev-docs	Developer documentation site (live)
posters-science-survey	Community survey on scientific poster sharing practices
poster-sharing-reuse-paper-code	Analysis code for the poster sharing and reuse study

Developers

How Poster Processing Works

When a user uploads a poster (PDF or image), the platform runs an automated extraction pipeline to convert the poster into structured, machine-readable metadata. Here is what happens:

flowchart TD
    A["User uploads poster\n(PDF or image)"] --> B["File stored securely"]
    B --> C{"File type?"}
    C -->|PDF| D["Text extraction\nvia pdfalto"]
    C -->|Image| E["Vision OCR\nvia Qwen2-VL"]
    D --> F["Llama 3.1 8B\n(optimized for poster extraction)"]
    E --> F
    F --> G["Structured JSON metadata\n(poster-json-schema, DataCite 4.7)"]
    G --> H["Metadata stored in database"]
    H --> I["Discoverable and searchable\non the platform"]
    G -.->|Optional| J["Deposit to Zenodo\nwith DOI"]

Loading

Text extraction. PDF posters are processed by pdfalto for layout-aware text extraction. Image posters (JPG, PNG) are processed by Qwen2-VL, a vision-language model that reads text directly from the image. This is handled by the extraction API.

Metadata structuring. The extracted raw text is structured into JSON by Llama 3.1 8B with parameters optimized for poster extraction. The output includes titles, authors, affiliations, content sections, and figure/table captions. See poster2json for the full extraction package.

What gets stored. The poster file is stored securely. Extracted metadata conforming to the poster-json-schema (aligned with DataCite 4.7) is stored in a PostgreSQL database and made searchable on the platform. Users can optionally deposit their poster to Zenodo for a persistent DOI.

Tech Stack

Layer	Technology
Frontend	Nuxt 3, Nuxt UI, Tailwind CSS
Backend	Nuxt server routes (Nitro)
Database	PostgreSQL via Prisma
Poster Extraction	poster2json (Python, Llama 3.1, Qwen2-VL, pdfalto)
File Storage	CDN-backed object storage
Repository Integration	Zenodo
Deployment	Docker

Getting Started

Prerequisites

• Node.js (22.x recommended)
• Yarn (1.x)
• Docker
• Volta (optional, for Node version management)

Setup

1. Clone the repository

   git clone https://github.com/fairdataihub/posters-science.git
   cd posters-science

2. Trust and install the required tool versions

   mise trust
   mise install

3. Install dependencies

   pnpm install

4. Add your environment variables

   cp .env.example .env

5. Start the development server

   pnpm dev

6. Open the application at http://localhost:3000

Development

Database

The application uses PostgreSQL. Run it locally with Docker:

docker-compose -f ./dev-docker-compose.yaml up -d

Stop the database:

docker-compose -f ./dev-docker-compose.yaml down

Prisma

The application uses Prisma to interact with the database.

UI

The application uses Nuxt UI for components and Tailwind CSS for styling.

For architecture details, see the Developer Documentation.

Citation

@software{posters_science2026,
  title = {Posters.science: A Platform for Sharing, Discovering, and Citing Scientific Posters},
  author = {Soundarajan, Sanjay and O'Neill, James and Portillo, Dorian and Patel, Bhavesh},
  year = {2026},
  url = {https://github.com/fairdataihub/posters-science}
}

Funding

This project is funded by The Navigation Fund (10.71707/rk36-9x79).

License

Distributed under the MIT License. See LICENSE for details.

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.