arXiv MCP Server

arXiv

Search, retrieve, and read scientific papers from arXiv.

Workflow

1. Search papers by keyword, category, author, date
2. Get metadata for a single paper or batch (up to 20)
3. Read full paper — returns the complete PDF for reading

Query Syntax

• all:transformer — search all fields
• ti:"attention is all you need" — title search
• au:"Vaswani" — author search
• cat:cs.CL — category filter
• abs:"reinforcement learning" — abstract search
• Combine with AND, OR, ANDNOT: cat:cs.AI AND ti:"large language model"
• Date range: submittedDate:[20240101 TO 20240331]

Common Categories

• cs.AI Artificial Intelligence, cs.CL Computation and Language (NLP), cs.CV Computer Vision
• cs.LG Machine Learning, cs.SE Software Engineering, cs.CR Cryptography and Security
• stat.ML Machine Learning (statistics), math.OC Optimization and Control
• eess.SP Signal Processing, q-bio Quantitative Biology, q-fin Quantitative Finance

Gotchas

• Rate limit: <= 1 request / 3 seconds per arXiv Terms of Use. Use the batch get tool instead of sequential single calls.
• Pagination: Use max_results (max 100) and start offset for large result sets.
• Read paper: Returns the PDF as a binary document. Not all very old papers have PDFs.
• Semantic Scholar: For citation graphs and references, use the Semantic Scholar bundle (separate provider).