Semantic Scholar

Semantic Scholar MCP Connector for Claude

A+

Search 200M+ academic papers with AI-powered TLDR summaries, influential citation tracking, and researcher profiles from the Allen Institute for AI.

4 tools Official Updated Jun 28, 2026 Official Vinkius Partner

Connect your AI agent to the world's most AI-enhanced academic knowledge graph, built and maintained by the Allen Institute for AI (AI2).

What you can do

  • AI-Powered Search — Find papers across 200M+ works with AI-generated TLDR summaries that distill each paper into a single sentence of key insight
  • Influential Citations — Beyond simple citation count, see how many influential citations a paper has received — those that meaningfully build upon the cited work
  • Multi-Format Lookup — Access papers by Semantic Scholar ID, DOI, ArXiv ID (arXiv:2106.09685), or PubMed ID (PMID:12345)
  • Citation Graph — Explore the full citation chain of any paper, with TLDR summaries for each citing work
  • Researcher Profiles — Find academics by name with paper counts, total citations, and h-index metrics

How it works

  1. Subscribe to this server
  2. Start searching 200M+ papers immediately — no API key required for basic usage

Who is this for?

  • AI/ML Researchers — find relevant papers in your field with instant TLDR summaries to quickly assess relevance before reading the full text
  • Graduate Students — build comprehensive literature reviews using the citation graph to trace how ideas evolve across the academic landscape
  • R&D Teams — evaluate researcher impact using influential citation counts and h-index metrics for talent scouting or collaboration decisions
academic-researchcitation-analysisknowledge-graphliterature-reviewscientific-dataresearch-profilessummarizationstem

4 tools expose this connector's capabilities to your AI agent.

get_semantic_citations

Essential for literature reviews and impact analysis. Find papers that cite a specific work on Semantic Scholar

get_semantic_paper

Accepts Semantic Scholar paper ID, DOI, ArXiv ID (e.g. arXiv:2106.09685), or PMID (e.g. PMID:12345). Get full paper details from Semantic Scholar by paper ID or DOI

search_semantic_author

Returns paper count, total citations, and h-index for each researcher. Find researchers and their publication metrics on Semantic Scholar

search_semantic_scholar

Returns papers with AI-generated TLDR summaries, citation counts, influential citation counts, and fields of study. Covers Computer Science, Medicine, Biology, Physics, and all STEM fields. Search 200M+ academic papers with AI-powered TLDR summaries and influence scores

See how to talk to your AI agent using Semantic Scholar.

What are the most cited papers on transformer architecture in deep learning?

Top results include: 'Attention Is All You Need' by Vaswani et al. (2017) — 🤖 TLDR: The Transformer, a new network architecture based solely on attention mechanisms, achieves superior quality on machine translation tasks. Cited 120,000+ times (25,000+ influential). Fields: Computer Science, Linguistics.

Get the full details of the LoRA paper using its ArXiv ID arXiv:2106.09685.

Found: 'LoRA: Low-Rank Adaptation of Large Language Models' by Edward Hu et al. (2021). 🤖 TLDR: LoRA freezes the pre-trained model weights and injects trainable rank decomposition matrices, reducing trainable parameters by 10,000x and GPU memory by 3x. Citations: 8,500+ (2,100 influential). Fields: CS, Mathematics. Available at arxiv.org/abs/2106.09685.

Find the researcher Yann LeCun and show me his publication metrics.

Found researcher Yann LeCun — Author at Meta AI / NYU. Papers: 950+, Total Citations: 380,000+, h-index: 162. Known for pioneering work in convolutional neural networks, self-supervised learning, and modern AI architectures. Profile link and full publication list available.

No API key is required for basic usage. The public API provides 5,000 requests per 5 minutes shared among unauthenticated users. For higher throughput, academic and institutional users can request a free API key at semanticscholar.org, which grants dedicated rate limits of 1–10 requests per second depending on the endpoint.

Related Connectors