Most of scientific paper sources provide some basic metadata about the paper (authors, publish date, DOI, etc) and sometimes the abstract and a link to the PDF of the paper. However, converting PDF to some structured data format like XML is a very tedious method,error prone and not scalable. After some research, I found some sources like CORE (https://core.ac.uk/services#dataset), Elsevier (https://dev.elsevier.com/sc_apis.html) and Springer (https://dev.springernature.com/) which provide the full text of the paper in XML or HTML format which is very usefull for data mining processes.

Are there any other scholarly sources which provide the full text of the papers in a format like XML or HTML ?

  • This may be related stackoverflow.com/q/10507049/2327328, although your question with full text will be harder since many papers are paywalled.
    – philshem
    Nov 23, 2018 at 7:23
  • Can you please add links to documentation about the sources you found? So that people can use them if needed.
    – philshem
    Nov 23, 2018 at 7:24
  • Also, see especially the CrossRef answer stackoverflow.com/a/38004100/2327328
    – philshem
    Nov 23, 2018 at 7:26
  • 1
    I added the links for the soources I mentioned
    – Agelos
    Nov 24, 2018 at 10:48

1 Answer 1


Three datasets come to my mind.


I found a fourth one, called unarXive. It is based on more than a million arXiv pre-prints, it is freely available, and contains rich metadata.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.