Contribute a paper

This is a non-commercial personal project. You parse the paper on your own machine with your own parser key, and submit only the resulting text — no PDF is uploaded and no key ever leaves your computer.

• Every submission is reviewed by a human before anything enters the database. Nothing is ingested automatically.
• You get a receipt code to check status. There is no account and no login.
• Please only submit papers you are allowed to share the text of.

Not sure what to submit? Browse papers that need help →

These are already in the database but under-extracted — a full parse of any of them adds more than a fresh paper would. Picking one fills in the form.

1. Get your free LlamaParse key Sign up at cloud.llamaindex.ai and create an API key. The free tier gives you 10,000 pages a month — a research paper is usually 10–30 pages, so that is hundreds of papers. Use the EU region if you are in Europe; a key minted in one region is rejected by the other.
2. Run this to extract the text Install the client with pip install llama-parse, drop your key and the PDF path into the snippet, and run it. The output is what goes in the form.
   Copy

   from llama_parse import LlamaParse
   
   parser = LlamaParse(
       api_key="YOUR_KEY",
       result_type="text",
       # EU keys only — delete this line if your key is US-region:
       base_url="https://api.cloud.eu.llamaindex.ai",
   )
   
   docs = parser.load_data("your_paper.pdf")
   text = "\n\n".join(d.text for d in docs)
   print(text)  # paste this into the form below

   Prefer not to run Python? Uploading the PDF in the LlamaParse web UI and copying the parsed text works just as well, as does any parser that returns clean text. Either way the parsing happens on your machine and your account — your key never touches this server, and no PDF is ever uploaded here.

3. Paste the text into the form and submit Put the extracted text in the form below along with the paper's title and DOI, then submit. You get a receipt code to track it, and a human reads every submission before anything enters the database.