ScholarGate
Get started
Open by default

Built on open records.
Given back the same way.

A research tool is only as checkable as the data under it. ScholarGate names the catalogs it stands on, keeps the records on the page, and publishes its own abstract layer back out for anyone to audit.

Read a method page

Which sources build the library?

The library is not scraped off the open web. Method pages are assembled from records in 3 open catalogs: OpenAlex, Semantic Scholar and Crossref.

Open catalogs are a deliberate choice, not a budget one. Anyone can look up the same record we did, and nothing important rests on a catalog only we can see.

  • OpenAlex, Semantic Scholar and Crossref, named in the open.
  • Each method page carries the records it was built from.
  • No private catalog sits between you and the evidence.

What can you verify yourself?

Trust that cannot be checked is just branding. Every claim on this page has a route you can walk without our help, and each one starts somewhere you can already reach.

Start on the page. Every method page lists the records behind it, so checking a claim begins with reading rather than with asking us.
Look the record up. The catalogs are public, so a record taken off the page can be found again in OpenAlex, Semantic Scholar or Crossref without us.
Read how it is built. The methodology page explains how a page earns its place and what gets published, before you decide to trust a single one.
Three drawers of records, all tied to the same seal.
One record that carries its check, next to a book left closed.
One answer, threaded out to the many languages that ask for it.
8,100+
Method pages
46
Interface languages
3
Open data sources

What happens when the library does not know?

A missing source is the one thing a research tool must never paper over. When the library has nothing to ground an answer on, it says so and stops there.

That rule costs us answers, and it is worth the cost. An invented reference looks exactly like a real one until a reviewer opens it, and by then it carries your name, not ours.

why the product is built this way

  • A gap is reported as a gap, never filled with a plausible citation.
  • Answers stay tied to the method pages they stand on.
  • Silence about a gap is treated as the expensive kind of error.

Open by default, or closed by default?

Open is not a badge in a footer. It decides what a reader is allowed to do with a record once they have found it, and that is the part of the argument that has to survive contact with a reviewer.

  • Look a record up yourself
    Open by default
    Included
    A closed index
    Not included
  • Read a page without an account
    Open by default
    Included
    A closed index
    Not included
  • See what a page was built from
    Open by default
    Included
    A closed index
    Not included
  • Reuse the layer in your own work
    Open by default
    CC-BY-4.0
    A closed index
    Not included
  • Cite the exact version you used
    Open by default
    A minted DOI
    A closed index
    Not included

Who can read the library?

Open data that only reads in one language is half open. The ScholarGate interface reads in 46 languages, so the catalog is not gated behind the language you happened to be taught in.

Method names travel badly between languages. A reader who knows a method under its local name should still land on the same page as everyone else.

  • The interface reads in 46 languages.
  • Method names and their aliases are searchable together.
  • The underlying records stay the same in every language.

How open is the library itself?

Taking from open data and giving nothing back is a quiet kind of enclosure. The abstract layer of every published method page, the definition, when to use it, strengths and limitations, and the records behind them, ships as a dataset under CC-BY-4.0 with a minted DOI.

Openness deserves an honest edge. The abstract layer is open; the deeper mechanics stay on the page. We would rather show you where the line runs than let you assume there is none.

  • Licensed CC-BY-4.0, so reuse asks for attribution and nothing more.
  • Minted with a DOI, so the version you cite stays citable.
  • A read only endpoint streams the same records, one method per line.

The dataset record

License
CC-BY-4.0
DOI
10.5281/zenodo.20539026
Bulk export
/api/v1/export
Scope
Abstract layer of published method pages
Open the dataset record

Questions about the data

Open scholarly catalogs: OpenAlex, Semantic Scholar and Crossref. Method pages keep the records they were built from, so a page can be followed back to its source without going through us.

Yes. The abstract layer is published as a dataset with a DOI, and a read only endpoint streams the same records as newline separated JSON, one published method per line.

The deeper mechanics of a method page stay on the site. What ships in the dataset is exactly what an anonymous reader sees on the page, so nothing is shown to crawlers that a person would not get.

No. The records a page is built from are visible on the page itself, and the catalogs behind them are public. Checking our work never requires being our customer.

That is what the license is for. Attribution is the only condition, and citing the DOI is the cleanest way to give it, because it pins the exact version you used.

Check a claim before you trust it.

Ask the library a method question and follow the answer back to the pages and records it stands on.

Ask the library