Google Scholar API

Google, as a matter of long-standing, publicly stated policy, does not operate an official, sanctioned application programming interface for Google Scholar -- there is no documented endpoint, no API key registration process, and no rate-limited free or paid tier comparable to those offered by Crossref, ORCID, Scopus, or OpenAlex. "Google Scholar API" as a search term almost always refers to one of two distinct things instead: (1) a third-party commercial scraping service (e.g. SerpApi, Apify, Oxylabs) that queries Google Scholar's public web pages on a customer's behalf and returns the result as structured JSON, or (2) an open-source scraping library (e.g. the Python package "scholarly") that automates a browser or HTTP session against the same public pages with no official sanction from Google. Both categories work by parsing Google Scholar's HTML search-results and profile pages, not by calling an endpoint Google built or documents for that purpose, and both are constrained by Google Scholar's robots.txt disallow rules and general Google Terms of Service language prohibiting automated querying without prior permission.

ByCASRAI Editorial Board

· Last updated 17 Jul 2026

Worked examples

Is an instance

An institutional research office wants to pull citation counts for 400 faculty profiles into a dashboard. Because no official Google Scholar API exists, the office either pays a third-party scraping service (e.g. SerpApi) for structured results scraped from Google Scholar's public profile pages, or accepts the throttling/CAPTCHA risk of running an open-source scraper like the Python package "scholarly" directly -- in both cases, 'the Google Scholar API' in a vendor's product name or a developer's code comment refers to this scraping layer, not an endpoint Google itself built or supports.

Is an instance

A developer installs the "scholarly" Python package and writes a script to bulk-download author profiles for a bibliometrics project. After a burst of a few hundred requests, Google Scholar starts returning CAPTCHA challenges instead of results -- the expected behavior when querying at a volume Google Scholar's anti-automation defenses are designed to catch, since the library is automating access to public search-results pages rather than calling a sanctioned, rate-limited interface.

Looks similar, but isn't

Not an instance

Calling the ORCID Public API, the Crossref REST API, the OpenAlex API, or the Semantic Scholar Academic Graph API to retrieve bibliographic or citation data is not 'using the Google Scholar API' -- these are genuinely official, documented, Terms-of-Service-sanctioned interfaces operated by their respective organizations, a fundamentally different arrangement from a third party scraping Google Scholar's web pages.

Not an instance

Manually looking up an author's Google Scholar profile in a browser, or clicking 'Export' to download a single citation in BibTeX format, is not API access at all -- it is normal, permitted interactive use of the public website, with none of the automation, scale, or Terms-of-Service exposure that 'API' implies.

There is no official Google Scholar API. Unlike Crossref, ORCID, Scopus, Web of Science, or OpenAlex, Google has never published a documented, sanctioned interface for programmatic access to Google Scholar’s search index, citation counts, or author-profile data — no API key registration page, no endpoint documentation, no published rate limits or pricing tier. This is a genuinely important, non-obvious fact for anyone trying to build a tool, dashboard, or automated report around Google Scholar data, because the phrase “Google Scholar API” is nonetheless in constant use across developer tooling, vendor marketing, and academic-support documentation — almost always referring to something else entirely: a third-party scraping service or library that queries Google Scholar’s public web pages without Google’s sanction.

Why there is no official API

Google has not published a single, authoritative explanation, but the pattern is consistent and long-standing (Google Scholar launched in 2004 and has never offered one): Google Scholar aggregates metadata and full-text indexing from a very large number of third-party publishers, repositories, and preprint servers under informal indexing arrangements, many of which come with restrictions on redistributing the underlying data at scale. An open, high-volume API would make it straightforward to bulk-extract that aggregated data far beyond what those underlying sources permit, and would also expose Google Scholar’s ranking and citation-matching methodology — which, unlike Scopus’s or Web of Science’s more disclosed indexing criteria, has never been fully documented — to systematic reverse-engineering. Google Scholar’s own robots.txt reflects this stance directly: it disallows crawler access to /scholar and /search, and disallows most /citations? parameters except a narrow, curated set of profile-listing views. Google’s general Terms of Service separately state that automated queries to any Google service require Google’s express advance permission, which Google Scholar has not granted for bulk querying.

What “Google Scholar API” actually refers to in practice

Because there is real demand for programmatic Google Scholar data — citation counts for a tenure dossier, h-index tracking for a research-office dashboard, bulk profile lookups for a bibliometrics study — and no sanctioned way to get it, two categories of workaround have filled the gap, and both are what people mean when they say “Google Scholar API”:

Commercial scraping-as-a-service products. Vendors such as SerpApi, Apify, Oxylabs, and several similarly positioned services sell access to an endpoint they operate, which in turn scrapes Google Scholar’s public pages (search results, citation pages, author profiles) on the customer’s behalf and returns the parsed result as structured JSON. These products market themselves explicitly as a “Google Scholar API” — the API is real and documented, but it is the vendor’s own scraping infrastructure, not anything Google built, operates, or has endorsed.
Open-source scraping libraries. Tools such as the Python package scholarly (and older, now largely unmaintained tools like scholar.py) automate an HTTP or browser session against Google Scholar’s public pages directly, with no vendor layer and no official sanction. These are free to use but carry the most direct exposure to CAPTCHA challenges and IP blocking, since there is no commercial intermediary managing that risk on the user’s behalf.

Both categories work by parsing HTML meant for a human browser, not by calling a documented, versioned interface Google maintains for machine consumption — the defining difference from a genuine official API.

Terms of Service and practical risk

Google’s general Terms of Service prohibit accessing Google services “through the use of any automated means (such as robots, spiders, or scrapers)” without Google’s express advance permission, and separately prohibit sending automated queries of any sort without such permission. Google Scholar’s robots.txt operationalizes this for crawlers by disallowing the search and most citation-related paths. In practice, Google enforces this less through legal action against individual researchers and more through automated technical defenses: request-rate monitoring, CAPTCHA challenges, and temporary or longer IP blocks once query volume crosses an undocumented threshold — something both scraping-service vendors and open-source-library users report running into at scale. Anyone building an institutional workflow around Google Scholar data should treat this as an ongoing operational risk, not a one-time setup decision: a scraping-dependent pipeline can degrade or break without warning if Google Scholar changes its page structure or tightens its anti-automation defenses, in a way a genuine, versioned API would give advance notice of.

Why institutions typically use other APIs instead

For programmatic bibliometrics and research-information-system work specifically, research offices and CRIS administrators generally reach for a database that actually offers a sanctioned, documented API rather than building on top of Google Scholar scraping:

Scopus (Elsevier) and Web of Science (Clarivate) — see Scopus vs. Web of Science vs. OpenAlex — offer subscription-based, documented APIs with defined coverage criteria, commonly licensed by university libraries and used to feed institutional CRIS platforms.
OpenAlex — a free, fully open, CC0-licensed catalog with a documented REST API and full bulk-data downloads, maintained by the nonprofit OurResearch, positioned explicitly as an open successor to the retired Microsoft Academic Graph.
Semantic Scholar — Ai2’s free Academic Graph API, with a public dataset release under an open license.
Crossref’s REST API and the ORCID API — for DOI-level metadata and author-record data specifically, both genuinely official, documented, and free at the public tier.

Each of these is a real, sanctioned API with published documentation, versioning, and (for most) a defined free tier — the category Google Scholar has no equivalent to. For an institution weighing which source to build an automated bibliometrics workflow on, “does this database offer an official API” is frequently the deciding factor, ahead of raw coverage size, precisely because a Google Scholar dependency means building on scraping infrastructure that can break or be blocked without notice.

Google Scholar API vs. Google Scholar Citations profile

These are two different things that are easy to conflate. A Google Scholar Citations profile is the free, self-managed author page a researcher creates on scholar.google.com, which Google Scholar itself computes citation-based metrics for (total citations, h-index, i10-index) and displays publicly. That profile can be viewed by anyone in a browser, and its public listing views are among the narrow set of paths Google Scholar’s robots.txt does permit crawling. “Google Scholar API,” by contrast, refers to attempts at bulk, automated, machine-readable access to that data (and Google Scholar’s underlying search index) at a scale beyond what viewing individual profile pages in a browser involves — the profile is the data source; the (unofficial) API is one way, among several, that people try to extract it programmatically.

Related terms

Google Scholar Citations profile
Scopus
ORCID API
h-index
OpenAlex: What It Is and How It Works (guide)
Semantic Scholar: What It Is and How It Works (guide)
Scopus vs. Web of Science vs. OpenAlex (comparison)

Use in your systems

JATS XML <role> element

xml

<role vocab="credit"
      vocab-identifier="https://casrai.org/dictionary/"
      vocab-term="Google Scholar API"
      vocab-term-identifier="https://casrai.org/dictionary/term/google-scholar-api" />

Schema.org DefinedTerm (JSON-LD)

json

{
  "@context": "https://schema.org",
  "@type": "DefinedTerm",
  "@id": "https://casrai.org/dictionary/term/google-scholar-api",
  "name": "Google Scholar API",
  "identifier": "https://casrai.org/dictionary/term/google-scholar-api",
  "description": "Google, as a matter of long-standing, publicly stated policy, does not operate an official, sanctioned application programming interface for Google Scholar -- there is no documented endpoint, no API key registration process, and no rate-limited free or paid tier comparable to those offered by Crossref, ORCID, Scopus, or OpenAlex. \"Google Scholar API\" as a search term almost always refers to one of two distinct things instead: (1) a third-party commercial scraping service (e.g. SerpApi, Apify, Oxylabs) that queries Google Scholar's public web pages on a customer's behalf and returns the result as structured JSON, or (2) an open-source scraping library (e.g. the Python package \"scholarly\") that automates a browser or HTTP session against the same public pages with no official sanction from Google. Both categories work by parsing Google Scholar's HTML search-results and profile pages, not by calling an endpoint Google built or documents for that purpose, and both are constrained by Google Scholar's robots.txt disallow rules and general Google Terms of Service language prohibiting automated querying without prior permission.",
  "inDefinedTermSet": "https://casrai.org/dictionary/domain/research-info-systems#set",
  "url": "https://casrai.org/dictionary/term/google-scholar-api",
  "sameAs": [],
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "publisher": {
    "@id": "https://casrai.org/#organization"
  },
  "dateModified": "2026-07-17T07:53:50",
  "inLanguage": "en"
}