Welcome to the Research Paper Relevance Checker — a simple and clean tool designed to save you hours of reading through endless research papers!
Instead of manually skimming pages just to figure out if a paper is relevant to your thesis, this tool turns the process into a fast, automated experience.
🌐 Try it live: Research Paper Relevance Checker
- Upload a research paper (PDF)
- Upload either your thesis topic or a draft document
- Get instant output:
- 🔍 Relevance score between your topic and the research paper
- 📖 Top citations extracted from the paper (filtered and cleaned)
- ✍️ Suggested literature review paragraph you can use as inspiration
- 📥 Download the full analysis as a PDF report for easy safekeeping!
- Cosine Similarity: Calculates the semantic similarity between your thesis and the research paper using their vector embeddings
- Cohere API:
- Embeds text into semantic vectors
- Summarizes the research paper
- Generates the literature review paragraph
- pdfminer.six: Extracts text from uploaded PDF files
- ReportLab: Dynamically generates the downloadable PDF report
- Streamlit: Builds the interactive web app UI
- Clean UI for uploading documents
- Fully interactive web app (built with Streamlit)
- Intelligent filtering of citations (removes meaningless year-only citations)
- Ability to download all outputs into a polished PDF
- Warning when a paper may not be relevant (relevance score under 30%)
If you want to run it locally:
pip install streamlit cohere numpy pdfminer.six reportlabThen:
streamlit run app.pyYou'll also need a Cohere API key — you can get one by signing up at https://dashboard.cohere.com/.
This tool is intended for research support purposes only. The relevance score and generated paragraphs are suggestions and should be critically evaluated in the context of your academic work.
Built by Sadia Khan ✨
Feel free to connect on LinkedIn or view more projects on GitHub!