A command-line tool for searching arXiv, downloading papers, and using Google's Gemini AI to analyze and extract information from research papers.

• Search arXiv for papers with customizable sorting and filtering
• Download PDFs of papers
• Ask questions about papers using Google's Gemini AI
• Summarize papers in different styles
• Extract structured data from papers
• Compare multiple papers
• Find related work using intelligent keyword extraction with Gemini and Google Scholar (via Serper API)
• Export citations in various formats (BibTeX, APA, MLA, Chicago, IEEE)

1. Clone this repository
2. Install dependencies:

   pip install -r requirements.txt
   

3. Set up API keys:
   • Copy .env.example to .env
   • Add your Gemini API key (get from https://ai.google.dev/)
   • Add your Serper API key (get from https://serper.dev/) if you want to use the related work feature

Run the app in interactive mode:

python main.py

Or use command-line arguments for non-interactive mode:

python main.py --query "your search query" --download 1 --ask 1 "What is the main contribution of this paper?"

• q - Enter a new arXiv search query
• n - Fetch the next page of results
• download [N,M,...] - Download PDF(s) for result number(s) N,M,... (comma-separated or single number)
• ask [N] "[Q]" - Ask Gemini a question about PDF N
• ask_fig [N] "[Q]" - Ask Gemini about figures/tables in PDF N
• sum [N] [style] - Summarize PDF N using Gemini (styles: simple, technical, key_findings, eli5)
• ext [N] [type] - Extract structured data from PDF N (types: methods, conclusion, datasets)
• compare [N1,N2,...] [type] - Compare multiple papers using Gemini
• rel [N] - Find related work for paper N using Google Scholar (requires Serper API key)
• cite [N] [format] - Export citation for paper N in specified format (formats: bibtex, apa, mla, chicago, ieee)
• set max [N] - Set max results per page
• set sort [field] [order] - Set sort order
• set model [name] - Set Gemini model
• show downloads - List downloaded/uploaded PDFs
• show model - Show current Gemini model
• help - Show help message
• quit - Exit

MIT