Reading each paper and copying its data into a table by hand is the slowest part of a literature review. Upload up to 5 PDFs, or pick them from My Library, and Paper Digest extracts the data into a table for you: one paper per row, one question per column. Start from ready-made columns - summary, study design, sample size, main findings, population, methods, limitations - or add your own in plain language, such as "Which dataset was used?" Each row is labeled with the paper's title and authors.
AI data extraction is only useful if you can check it. Each cell in the table comes with a quote copied from the paper - click the cell to see that passage highlighted in the source text. If a paper does not report something, the cell says Not reported instead of guessing, and if a quote cannot be found in the paper, the cell is flagged with a warning sign so you know to look closer. Answers can be written in 12 languages; quotes stay in the paper's own language.
Pilot your data extraction form: define the fields as columns, test them on a few included studies and refine the wording before extracting at scale. Columns can hold numbers, such as sample sizes, or descriptions, such as methods and limitations. Export the data extraction table to CSV or copy it into Excel or Google Sheets. PDFs with selectable text and TXT files work; scanned PDFs are not supported yet. Need papers first? Start with a Literature Review or Deep Research report.