Skip to content
SolvesEverything
← All postdoc tools
TOOL 19 · Lab & Bench

Paper Explorer

Enter a PubMed ID, PMCID or DOI — or drop in a PDF — and read a paper in minutes: what was changed and what moved, figure by figure, with every claim quoted. Get the concept map, reagents and accessions, citations and a journal club pack — or compare papers and triage a literature search.

Rule-based · extractive · no AI · PDFs never leave your browser

Loading this browser tool…

How to use it

One paper. Type a PubMed ID (e.g. 23390586), a PMCID or a DOI and press “Fetch and analyse”, or drop in a PDF. Open-access papers are analysed from their full text in Europe PMC; for other papers only the abstract is available, so upload your own PDF for a full analysis. Start with Summary: the research question, the central sentences, Results at a glance (what was changed, what was measured and which way it moved) and the key findings. Then walk through Figures, the Concept map, the highlighted full text in Read, Reagents & resources (antibodies, strains, accessions, software, doses), Literature (who cites the paper and its most-cited references) and the Journal club outline. Click any underlined name to see every sentence about it.

Share. A paper opened by identifier gets a link (/tools/paper-explorer/?id=23390586) that opens the same analysis for anyone. Compare puts two to five papers side by side — shared entities, findings on the same readout and where they differ. Literature triage turns a Europe PMC search or a list of IDs into one table of each paper's main finding, entities and models.

How it works

Paper Explorer does not use an AI model and does not write new text. It splits the paper into sections and sentences and labels each sentence as a new finding (this paper's own data), background (prior work, usually cited), interpretation, study aim or method. It ranks sentences with a TextRank-style similarity graph, recognises biomedical entities with a curated vocabulary completed by Europe PMC's text-mined terms and abbreviations defined in the paper, and extracts statistics from the text and figure legends. Results at a glance reads perturbations (“Irg1-KO”, “loss of”, “treated with”), readouts and direction words (increased, reduced, no difference) from new-finding sentences only; loss-of-function results also become inferred relations in the concept map (“X-KO → more Y” means X restrains Y), marked as such.

Limits

Rules miss things and occasionally misread them: every row, relation and resource links to the sentence it came from, so check it against the paper. Text inside figures, images and most PDF tables is not visible, scanned PDFs cannot be read, and methods kept in a supplement are not part of the open-access text. Discussion prompts in the journal club pack are templates, not the authors' words. Entity recognition favours immunology and metabolism. Quotations belong to the original authors; the tool is for reading and note-taking, and nothing is republished or stored.

Privacy

Identifiers and search terms you type are sent to Europe PMC (EMBL-EBI) to retrieve papers, text-mined terms and citations; uploaded PDFs are processed only in this tab. Nothing is saved unless you export it.

Privacy and local-data controls · Credits and licenses · Interpretation limits