CATO

Library and Deep Search

Search the scholarly literature with a research question, save the papers you want, extract evidence, and use selected papers in your analysis and writing.

A search can miss papers. Coverage, indexing delays, and access rights all affect the results. Keep your question, and run your own database searches for a formal review.

Where It Is

Select Library in the project sidebar. The tabs at the top are Library, for the papers you save, Deep Search, for searching, and Saved reports, for the documents CATO writes. All of it stays with the project.

Deep Search

Deep Search returns up to 100 papers for a research question. The box under the search field says which kind of search your account runs.

Most accounts. Your question goes as you wrote it to an index of scholarly publications, which ranks results by how closely they match its meaning. Boolean operators and quotes are not interpreted, and nothing is screened out, so read the whole list.

Organization accounts with restricted sources. CATO writes a query for each of PubMed, Europe PMC, Crossref, OpenAlex, arXiv, bioRxiv and medRxiv, and eligibility criteria from your question. It screens the best-matching papers on title and abstract and shows only the papers it includes. Screening is automatic and reads abstracts only, so it does not replace your own screening or a documented database search.

  1. Write the research question as a sentence, naming the population, intervention or exposure, comparator, and outcomes where they matter.
  2. Narrow the publication years if you want to.
  3. Run the search, read down the results, and save the papers you want to the Library.

Example

“What disease-free and overall survival do adjuvant immune checkpoint inhibitors achieve, compared with chemotherapy or observation, in adults with resected stage II–III non-small-cell lung cancer?”

For ClinicalTrials.gov registrations and patents, ask in Chat. Chat can also search the literature during an analysis or writing task, and can run a Boolean strategy in PubMed and report the count.

Reading the Results

The order shows how well a paper matches your question, not how good the study is. A preprint and the journal article it became appear as one result.

  • Open a record to check its title, authors, journal, year, DOI or other ID, abstract, and links.
  • Save the papers you want to keep, because a later search can return a different order.
  • For a formal review, keep your own record of why papers were excluded.

Full Text

When a legal open-access copy exists, for example in PMC or Europe PMC, CATO can fetch it and you can read and comment on it in the Workspace. When it cannot, the record says that only the abstract or citation details are available.

Know what you have: an abstract cannot support detailed claims about methods, subgroups, adverse events, adjustments, or limitations that appear only in the full text.

Library and Collections

The Library holds the papers you chose to save. Sort them into collections, add tags and notes, flag them for review, and open whatever text is available.

  • Use collections for stages such as Candidate, Included, Excluded, Background, and Methods.
  • Keep your reasons for excluding a paper separate from informal notes if a review will rely on them.
  • Fix duplicate or conflicting citation details before you export citations or write a document.

Systematic Review

The Systematic Review tab on the Deep Search page is available on Pro, Pay-as-you-go, and Enterprise. It pulls the same fields from each paper you select, such as population, intervention or exposure, comparator, outcomes, design, sample size, effects, and limitations, then writes a report in which each claim cites its papers. Reports are kept on the Saved reports tab.

  • Run it on a set of papers you have already chosen, not on every search result.
  • Check the fields that matter against the paper itself.
  • Tell “not reported” apart from “not found in the text CATO had”.

It does not make a review systematic by itself. A systematic review still needs a protocol, documented searches, deduplication, independent screening, recorded exclusions, checked extraction, a risk-of-bias assessment, and reporting to the applicable standard.

Citation Network

Network, beside List on the Deep Search results, draws how the papers cite each other. It can show clusters, influential papers, and papers that link two areas. Coverage varies by source and era, and a much-cited paper is not necessarily a good one.

Watches

A watch checks for new papers on a topic. Set one up under Alerts in the Library sidebar. Watches check while the Library or Deep Search page of that project is open, at most once a day; they do not check while those pages are closed.

PlanActive watches
Standard10
Pro50
Pay-as-you-goNo fixed limit
EnterpriseNo fixed limit

A new paper can take time to be indexed, so it may not appear even when a watch has just checked.

Send Papers to Chat

  1. Select papers in the results or the Library.
  2. Select Send to Chat.
  3. Choose Ask with context, Analyze and extract, Draft with citations, or Plan next experiments.
  4. Read the prompt it fills in, and change it or add your question.
  5. In Chat, open the cited papers before you accept the answer.

Say “use only the attached sources” when that matters, and ask CATO to flag claims the available text does not support.

Saved Reports and Exports

Reviews and other generated documents are listed newest first on the Saved reports tab. Open one to read it, or delete it from its page. From the Library, you can export citations as BibTeX, RIS, EndNote XML, CSL JSON, or CSV.

Working with Covidence

Covidence has no public API, so studies move between Covidence and CATO as files.

  1. In Covidence, choose Export to file, then References, the stage and the list (for example Included), and download it as CSV or RIS. The CSV keeps the Covidence #, study label, notes and tags; the RIS file does not. Extracted data downloads as CSV.
  2. Upload the file to the project and ask CATO to read it. CATO builds a table of the studies, with DOIs and PMIDs, exclusion reasons from an excluded list, and likely duplicates flagged, and saves it next to your file in Files.
  3. Ask CATO to fetch the full texts. It reads open-access papers in full and works from the abstract for the others.
  4. To send studies to Covidence, ask CATO to export them for Covidence. It saves an RIS file in the exports folder in Files, with missing abstracts looked up in PubMed. Import that file in Covidence at the stage you choose.
  5. CATO can also write extracted data as a CSV in Covidence's Extraction 2 layout. Covidence cannot import extracted data, so this file is for your records and analysis.

While It Runs

Searches, extractions, and reports show progress and warnings as they run, and they keep going if you leave: come back to the page for the finished results. Read any warnings before you draw conclusions from a short list.

Search and Review Checklist

  • Save the exact question, date, and year range.
  • Check indexing dates, spelling variants, controlled vocabulary, and identifiers.
  • Look for duplicate versions, retractions, and corrections.
  • Check eligibility, extraction, and risk-of-bias judgments against the full text.
  • Confirm every generated citation and quotation in the source.
  • Record where you had only an abstract or citation details.
  • Download the search and evidence files you need to reproduce the work.