Read, annotate, and talk to your research papers—effortlessly.
PaperStack is an open-source AI-driven research assistant that helps you organize, annotate, and understand academic papers with ease. From intelligent metadata extraction to interactive paper-specific chat, PaperStack transforms your research workflow into a seamless, searchable, and smart experience—all starting with a single PDF upload.
- Metadata Extraction: Extracts title, authors, abstract, keywords, and affiliations using GROBID.
- Organized Library: Upload, manage, and group research papers by tags, topics, or folders.
- Smart PDF Annotation: Read and highlight papers in-browser using @react-pdf-viewer.
- Chat with Your Paper: Ask context-aware questions with RAG-powered responses.
- Keyword Standardization: Maintains consistent tagging for improved search and grouping.
- Custom Upload Workflow: Upload PDFs, extract metadata, and start interacting in a single flow.
# 1. Clone the repository
git clone https://github.com/ankraj1234/PaperStack.git
cd paperstack
# 2. Start GROBID (for metadata extraction)
cd grobid/
./gradlew run
# GROBID will be available at http://localhost:8070
# 3. Start ans setup the backend (in the project root)
pip install -r backend/requirements.txt
uvicorn backend.app.main:app --reload
# Ensure your .env file is configured
# 4. Start the frontend
cd frontend
npm install # Run only once
npm start
# App runs at http://localhost:3000- Frontend: React, CSS, @react-pdf-viewer
- Backend: FastAPI (Python), SQLAlchemy
- Database: MySQL
- PDF Parsing: GROBID
- Vector Store: FAISS