Getting full text, legally

Automatic where it is allowed

Full text is retrieved only by legal routes: open-access copies, open repositories, and publisher services your organisation is licensed for. A copy that is fetched is stored with the source it came from and its licence.

For each paper, the platform records whether an open copy exists, where it is, under which licence, and when this was last checked.

One status for every paper

A paper shows the same full-text status in the Library, on the Full text page and on the PDFs page. Each status comes with one sentence saying what it means and, where there is one, the next thing you can do.

• Attached: a PDF of this paper is attached. Choose Read to open it.
• Available: an open copy was found and can be fetched. Run Find full text.
• Downloading: the current run is working on it.
• Library lookup: your institution’s library link is being checked.
• Institutional access required: open the link, sign in with your own account in your own browser, then upload the PDF. Nothing is fetched with your sign-in.
• Manual upload needed: get the PDF from the publisher, an interlibrary loan or the authors, and upload it.
• Unavailable: no legal copy was found.
• Policy blocked: a copy exists, but its licence or the provider’s terms do not allow automatic retrieval.
• Failed — retry: the last attempt did not finish. Retry starts a new attempt for that paper.
• Not looked for yet: nothing has looked for this paper so far.

Choosing which papers to look for

Find full text can run over every included paper without a PDF, over the papers the scorer marked as worth extracting, over the papers you ticked in the Library, or over a pasted list of DOIs and PubMed IDs (up to 500). A pasted identifier that is not in the library is listed for you to import first; no paper is created for it. Papers that already have a PDF are not fetched again.

A manual list for everything else

When no legal automatic route exists, the paper goes to a manual list with the reason. You obtain the PDF through your own library access and upload it.

The list separates papers that have a free copy you can download yourself from papers with no open copy, which need a subscription or an interlibrary loan. Sorting by score puts the most promising papers first, and a paper can be assigned to one person so that two people do not fetch the same file.

What the platform never does

The platform never bypasses a paywall, never asks for your university password, and does not route requests through a library proxy. No setting changes this, because a dataset built from copies obtained improperly would put the researcher at risk.

Supplementary files

Much functionality data is published in supplementary tables. Add supplementary files alongside the article so that they are stored and linked to the paper.

Values are not yet extracted automatically from supplementary files. They are kept for a person to read and enter.