Running a literature review
Sharly’s multi-document analysis makes literature reviews faster and more thorough. Here’s a typical workflow:1. Upload your papers
Drag and drop PDFs, Word documents, presentations, EPUB, HTML, text, images or audio into your workspace, up to 100 MB per file. Scanned PDFs are read with OCR. You can also add papers by arXiv ID or DOI, and keep them all in your Library. Each document can be up to about 250 pages on Free and about 1,250 pages on paid plans; see Limits.2. Summarize
Sharly generates a concise summary for each document, including bullet points and suggested follow-up questions (“icebreakers”) to guide your reading.3. Extract key findings
Use extraction to pull out methodologies, hypotheses, results, tables, and statistics from each paper. Sharly identifies key topics and entities automatically.4. Compare across sources
With multiple documents in one chat, ask Sharly to compare findings, methods or definitions across them.5. Cite and export
Every answer shows the passages it came from, with page numbers for PDFs. Use the Cite button to add a journal article to your reference list or reference manager. Export the chat as PDF, TXT or Markdown.Citations and academic integrity
Sharly links every generated insight to the exact passage in the original document. This source-verification approach supports academic integrity by letting you:- Verify claims — open any citation to read the passage in the original document.
- Cite the paper — copy a ready-made citation, or export it to your reference manager.
- See what a paper cites — Insights lists the references from each paper’s own bibliography, in APA and MLA.
Working with scanned PDFs
Older journal articles and archival documents often exist only as scanned PDFs. Sharly’s OCR engine processes these automatically, converting scanned text into searchable, summarizable content. OCR runs when a PDF has no text layer at all. A PDF that mixes born-digital pages with scans is read from its text layer only, so split or re-scan mixed documents before uploading. Scanned PDFs, TIFFs and images added from a link, including Google Drive and Dropbox links, are read the same way. Scanned documents can be up to 2,000 pages. For best results:- Use high-resolution scans (300 DPI or above)
- Ensure pages are not heavily skewed or cropped
- Multi-column layouts are supported
Tips for extracting research data
Use specific prompts to get structured results:- “Extract the methodology section from each paper” — pulls research design details across documents
- “List all statistical findings with p-values” — surfaces quantitative results for comparison
- “Compare the hypotheses across these papers” — identifies alignment or divergence in research questions
- “Summarize in 5 bullets” — quick overview when you need to triage a large reading list