Skip to main content
The Library (Dashboard → Library) works like a reference manager. On the left is the list of collections, in the middle the list of items, and on the right the details of whichever item you’ve selected.

Layout

Adding papers

1

Pick a collection

Choose the Collection for new files. Every file has to go into a collection.
2

Add the files

Drag PDFs anywhere onto the list, or use Add files. Each upload shows its progress.
3

Wait for processing

Each item’s status updates on its own. You can keep working while files process.
You can also bring papers in from Zotero or Google Drive.

Item statuses

Finding papers

  • Search matches title, author, year and DOI.
  • Filter by article type (research article, review, theoretical, policy, grey literature) and by status.

The item pane

Details

Title, authors, publication, year, volume, issue, pages, DOI, keywords and abstract. Edit anything that’s wrong. Good details give you better citations and exports.

Study record (extraction)

Fetchlake reads the paper and fills in a structured record. The fields depend on the article type. For a research article they include:
  • Study design, methods, setting, country, population and sample size
  • Intervention or exposure, and comparison
  • Outcomes: each with its measure, timepoint, result, effect size and significance
  • Main findings, limitations and a summary
Re-extract runs the AI again. It uses credits, so correct small mistakes by hand instead.
Extraction is a starting point, not a final answer. Always check effect sizes, sample sizes and significance against the paper before you rely on them.

The reader

Double-click an item (or use Read the document with its details and notes) to open the reader: the item’s details on the left, the PDF in the centre and your notes underneath. Notes save automatically.

Tips

  • Fill in article type first. It decides which extraction fields you see.
  • Sort the list by status and work through Needs review items in one go.
  • If a PDF won’t read, it’s usually a scan. Run it through OCR and upload it again.