yzxoi/paperly-codex/semantic-scholar

Semantic Scholar

Overview

Queries the Semantic Scholar Academic Graph API for paper metadata, citations, references, author profiles, and semantic search. High-coverage index of 200M+ papers across all scientific fields with structured, machine-readable data.

When to Use

• User wants paper metadata (abstract, authors, year, venue, citation count) by title or DOI
• User asks for papers citing or referenced by a specific work
• User wants to find papers by a specific author or lab
• Semantic search over a topic when PubMed/arXiv coverage is insufficient
• Need structured data for building citation graphs or lit review databases

Key Capabilities

• Search by keyword, title, author, or fieldsOfStudy
• Retrieve full citation and reference lists per paper
• Author profiles with h-index, citation count, affiliated institution
• Batch lookups by DOI, arXiv ID, or S2 paper ID
• Semantic similarity search (embedding-based, not just keyword)
• Open access PDF links where available
• TLDRs: auto-generated 1–2 sentence summaries per paper

Usage Examples

Search for papers on a topic

s2.search(
    query="few-shot learning protein structure prediction",
    fields=["title", "authors", "year", "citationCount", "tldr"],
    limit=20
)

Get papers citing a specific work

s2.get_citations(
    paper_id="arXiv:1706.03762",
    fields=["title", "year", "authors"],
    limit=50
)

Author profile lookup

s2.get_author(
    name="Yoshua Bengio",
    fields=["hIndex", "citationCount", "papers"]
)

Batch DOI lookup

s2.batch_lookup(
    ids=["DOI:10.1038/...", "DOI:10.1101/..."],
    fields=["title", "abstract", "tldr", "openAccessPdf"]
)

Output Format

Returns structured JSON with paper metadata. TLDRs are available for most papers published after 2018. OpenAccessPdf field provides direct PDF URL when available.

Best Paired With

• related_work_scout for metadata, citations, and reference chasing

Notes

• API key optional but recommended for higher rate limits (100 req/5min vs 10 req/5min)
• Register free at https://www.semanticscholar.org/product/api
• fieldsOfStudy filter supports: "Computer Science", "Biology", "Medicine", "Physics", etc.
• For bulk exports >10k papers, use the S2 Dataset releases instead of API