Use when searching academic papers, looking up citations, finding authors, or getting paper recommendations using the Semantic Scholar API. Triggers on queries about research papers, academic search, citation analysis, or literature discovery.
Use when searching academic papers, looking up citations, finding authors, or getting paper recommendations using the Semantic Scholar API. Triggers on queries about research papers, academic search, citation analysis, or literature discovery.
Requires python3 and the `requests` package. Set S2_API_KEY for higher rate limits (request at https://www.semanticscholar.org/product/api#api-key). Works unauthenticated with strict rate limits.
Search academic papers via the Semantic Scholar API using a structured 4-phase workflow.
Critical rule: NEVER make multiple sequential Bash calls for API requests. Always write ONE Python script that runs all searches, then execute it once. All rate limiting is handled inside s2.py automatically.
Phase 1: Understand & Plan
Parse the user's intent and choose a search strategy:
Decision Tree
Default to search_bulk(). Per Semantic Scholar's own docs, bulk search is preferred over relevance search for most cases because relevance search is more resource-intensive. Use search_relevance() only when you need TLDR fields or author/citation details inline.
User wants...
Strategy
Function
Broad topic exploration
Bulk search (preferred)
search_bulk() with build_bool_query()
Need TLDR / inline author details
Relevance search
search_relevance()
Precise technical terms, exact phrases
Bulk search with boolean operators
search_bulk() with build_bool_query()
Specific passages or methods
Snippet search
search_snippets()
Known paper by title
Title match
match_title()
Known paper by DOI/PMID/ArXiv
Direct lookup
get_paper()
Papers citing a known work
Citation traversal
get_citations()
Related to one paper
Single-seed recommendations
find_similar()
Related to multiple papers
Multi-seed recommendations
recommend()
Find a researcher
Author search
search_authors()
Researcher's profile
Author details
get_author()
Researcher's publications
Author papers
get_author_papers()
Query Construction Rules
Ambiguous terms (e.g., "stem cells" could mean mesenchymal or stem-like T cells): Use build_bool_query() with exact phrases and exclusions
Example: build_bool_query(phrases=["stem-like T cells"], required=["CD4", "TCF7"], excluded=["mesenchymal", "hematopoietic stem cell"])
Multi-context queries (e.g., "topic X in cancer AND autoimmunity"): Plan separate searches, deduplicate with deduplicate()
Broad topics: Use search_relevance() with filters (year, venue, fieldsOfStudy, minCitationCount)
Plan Filters
Filter
Use when
year="2020-"
Recent work only
publication_date="2024-01-01:2024-06-30"
Precise date range (YYYY-MM-DD)
fields_of_study="Medicine"
Restrict to domain
min_citations=10
Only established papers
pub_types="Review"
Find reviews/meta-analyses
pub_types="ClinicalTrial"
Clinical trials only
open_access=True
Only open access papers
Checkpoint: Before proceeding, verify: (1) search strategy matches user intent, (2) filters are appropriate, (3) query is specific enough to avoid irrelevant results.
Phase 2: Execute Search
Write ONE Python script that begins with the standard prelude below, then runs all searches:
# --- Standard prelude (use in every script) ---import sys, os, glob
_candidates = [
os.path.expanduser("~/.claude/skills/semanticscholar-skill"),
os.path.expanduser("~/.openclaw/skills/semanticscholar-skill"),
*glob.glob(os.path.expanduser("~/.claude/plugins/**/semanticscholar-skill"), recursive=True),
*glob.glob(os.path.expanduser("~/.codex/skills/semanticscholar-skill")),
".",
]
SKILL_DIR = next((p for p in _candidates if os.path.isfile(os.path.join(p, "s2.py"))), None)
if SKILL_DIR isNone:
raise RuntimeError("Cannot locate semanticscholar-skill (s2.py not found)")
sys.path.insert(0, SKILL_DIR)
from s2 import *
# --- end prelude ---# Build precise query
q = build_bool_query(
phrases=["stem-like T cells"],
required=["CD4", "IBD"],
excluded=["mesenchymal"]
)
papers = search_bulk(q, max_results=30, year="2018-", fields_of_study="Medicine")
papers = deduplicate(papers)
print(format_results(papers, "Stem-like CD4 T cells in IBD"))
Save to /tmp/s2_search.py, then run with python3 /tmp/s2_search.py in a single Bash call. Rate limiting, retries, and backoff are automatic inside s2.py.
No API key: The skill works without S2_API_KEY. When the key is absent or invalid, s2.py automatically switches to unauthenticated mode (no x-api-key header) and widens the request gap to 5 s. Per S2 docs, anonymous calls share a global 1000 req/s pool across all unauthenticated users and can be "further throttled during periods of heavy use" — so a conservative 5 s gap protects against the heavy-use throttling, even though the steady-state pool is generous. If you still see sustained 429s, raise _MIN_GAP to 10 s. Keep max_results ≤ 30 per search and combine fewer searches per script. S2 recommends including an API key on every request — get one at https://www.semanticscholar.org/product/api#api-key-form.
Checkpoint: Verify the script ran successfully (no exceptions) and returned results. If 0 results, broaden the query or relax filters before presenting.
Worked Examples
Each example below assumes the standard prelude from Phase 2 is at the top of the script.
Example 1: Author workflow — "Find papers by Yann LeCun on self-supervised learning"
authors = search_authors("Yann LeCun", max_results=5)
print(format_authors(authors))
# Use the first match's ID to get their papers
author_id = authors[0]["authorId"]
papers = get_author_papers(author_id, max_results=50)
# Filter locally for topic
ssl_papers = [p for p in papers if"self-supervised"in (p.get("title") or"").lower()]
print(format_results(ssl_papers, "Yann LeCun - Self-Supervised Learning"))
Example 2: Citation chain with intent — "Who cited the Transformer paper and how did they use it?"
Example 3: Multi-seed recommendations with BibTeX export — "Find papers like these two but not about NLP"
recs = recommend(
positive_ids=["DOI:10.1038/nature14539", "ARXIV:2010.11929"],
negative_ids=["ARXIV:1706.03762"],
limit=20
)
print(format_results(recs, "Vision papers like Deep Learning & ViT, excluding NLP"))
# Export BibTeX for top results
bib_data = batch_papers([r["paperId"] for r in recs[:10]], fields="title,citationStyles")
print(export_bibtex(bib_data))
Phase 3: Summarize & Present
Use format_results() for consistent output (summary table + top-10 details)
If user's language is Chinese, present summaries in Chinese
Always note total results count and search strategy used
Highlight most relevant papers based on the user's specific question
Phase 4: User Interaction Loop
After presenting results, always offer these options:
Translate — titles/summaries to Chinese (or other language)
Details — full abstract for specific paper numbers
Refine — narrow or expand search with different terms/filters
Similar — find papers similar to a specific result (find_similar())
Citations — who cited a specific paper and how (get_citations() + format_citations() for intent labels)
Export — save results via export_bibtex(), export_markdown(), or export_json()
Done — end search session
Loop until user says done. Each follow-up uses the same single-script pattern.
Use the standard prelude from Phase 2 at the top of every script. Then call any of the functions below — the module's docstring (help(s2) or read s2.py) lists each by phase with one-line summaries.
Minimize fields. Per the official S2 tutorial: "Avoid including more fields than you need, because that can slow down the response rate." Only add abstract, references, or citations when the user explicitly needs them.
sort Parameter Values (bulk search only)
The sort kwarg accepts only these three values:
Value
Meaning
citationCount:desc
Most-cited first (default)
publicationDate:desc
Newest first
paperId:asc
Stable deterministic order (useful for pagination)
Recommendations Limits
find_similar() and recommend() return at most 500 papers per call (limit max = 500).
Datasets API Functions
For bulk download of full S2 datasets (papers, authors, abstracts, embeddings, etc.):
Function
Purpose
Requires key?
list_releases()
List all available release date strings
No
list_datasets(release_id="latest")
List datasets in a release
No
get_dataset_links(release_id, dataset_name)
Pre-signed download URLs for a dataset
Yes
get_dataset_diffs(start, end, dataset_name)
Incremental diffs between two releases
Yes
Available dataset names (pass as dataset_name):
Name
Description
Approx size
papers
Core paper attributes (title, authors, date, etc.)
~200M records, 30 × 1.5 GB
abstracts
Paper abstract text where available
~100M records, 30 × 1.8 GB
authors
Author core attributes (name, affiliation, paper count)
—
citations
Citation relationships between papers
—
embeddings-specter_v1
Dense SPECTER vector embeddings of papers
~120M records, 30 × 28 GB
publication-venues
Venue metadata
—
s2orc
Full-body text from open-access PDFs
—
tldrs
Short natural-language summaries
~100M records, 30 × 200 MB
All datasets are delivered as JSON Lines (one record per line). The diffs response contains update_files (insert/replace by primary key) and delete_files (remove from dataset).
Rate Limiting
s2.py adapts automatically based on whether S2_API_KEY is set:
Mode
Gap
Official limit
Retries
Authenticated (valid key)
1.1 s
Introductory 1 req/s per key, dedicated quota, cumulative across all endpoints (raisable on request)
5× exponential backoff (2s→60s)
Unauthenticated (no key or invalid key)
5.0 s
1000 req/s shared globally across all anonymous users; "may be further throttled during periods of heavy use"
5× exponential backoff (2s→60s)
S2 recommends including an API key on every request, even for endpoints that work anonymously — it gives you a dedicated quota, a smoother experience under load, and better support if you need help. The introductory 1 req/s key can be raised on request. Get one at https://www.semanticscholar.org/product/api#api-key-form
The anonymous 1000 req/s pool is generous in steady state, but the docs explicitly warn it can be throttled hard during heavy use — that is why _MIN_GAP defaults to 5 s without a key, not the 1 ms a 1000 req/s budget would technically allow. If your workload still hits sustained 429s, set _MIN_GAP = 10.0 in s2.py or get a key. The 1 req/s key budget is cumulative across all endpoints, so chained calls (e.g. get_paper → get_citations) count separately.
Bulk Search Response Structure
search_bulk() returns a list of papers already unpaginated. Internally the raw response has:
Field
Type
Meaning
total
integer
Estimated total matching papers (not exact)
token
string
Present when more pages exist; pass in next request
data
array
Papers for this page
s2.py handles token pagination automatically — you only see the final flat list.
