The hardest part of switching tools is the empty library. Folio's on-ramp is built to get years of scattered research in quickly: your PDFs, your reference exports, and even the planning you've already done with an AI assistant.
Where to start
You'll find these in a few places, so the on-ramp is always within reach:
- New accounts see a Get started with Folio panel on the dashboard, with a Paste a ChatGPT / Claude thread shortcut and an Import button for sources, until you create your first document.
- New document in the sidebar has a Bring something in group: Import from AI chat and Import .docx.
- The Sources library → Add source has every importer, including Bulk PDFs.
Bulk-import a folder of PDFs
If your sources live as PDFs on your desktop or in a cloud drive, this is the fastest path.
- Open Add source → Bulk PDFs.
- Drag in a whole folder, or select many files at once (up to 50 per batch).
- Folio reads each one, matches it to verified metadata, and shows a list you can review.
- Pick the ones to keep and click Add to library (the button counts them for you).
For each PDF, Folio pulls the text, finds the DOI, and confirms the citation against CrossRef, so you get clean, verified metadata, not a guessed title. Papers without a discoverable DOI are still imported and flagged Check title so you can fix them in a second. Scanned PDFs of up to 25 pages are read with OCR and marked OCR'd. Anything already in your library is skipped automatically (matched by DOI or title), and the original PDFs are stored so you can read and annotate them in Folio.
Bulk PDF import uses no AI credits. It's free on every plan.
Import one PDF from Google Drive or OneDrive
Open Sources → Add source → Google Drive or OneDrive, choose a PDF, and Folio sends it through the same metadata, reference, and private-storage flow as a local PDF. Only the file you choose is shared with Folio, and the provider's access token is used for that request and not stored. If the card says the import isn't set up yet, download the file and use PDF Upload or Bulk PDFs instead.
Paste an AI research chat
Been planning a paper with ChatGPT or Claude? Don't start over.
- New document → Import from AI chat (or the dashboard panel → Paste a ChatGPT / Claude thread).
- Paste the conversation and click Analyze chat.
- Folio lifts out a working thesis, an outline, the papers you discussed, and the key claims. Review what it found, then choose Create draft.
- Folio creates an empty document and adds the sources you kept to your library. The thesis, outline and claims open beside the document in the Research brief panel, each with a copy button, so you write the draft yourself from them. Nothing from the chat is written into the page.
Because AI assistants routinely invent citations, Folio verifies every source against CrossRef before trusting it. Sources it can confirm are marked Verified and checked by default; ones it can't are flagged Unverified and left unchecked, so a hallucinated reference never sneaks into your library.
Get a cleaner import with the prompt
In the import window, click Copy prompt. Paste it into your ChatGPT or Claude chat first, and the assistant will reply with a tidy, labeled brief: topic, thesis, outline, sources with DOIs, and claims. Paste that reply back into Folio. Because it includes DOIs and is told not to invent sources, more of your citations come back Verified.
AI chat import is metered: the free Scholar plan includes 2 a month, Fellow gets 10 a day, and Chair 20 a day.
Import a Word document, or a whole folder of them
Have your manuscript in .docx? Folio brings the text, headings, tables, footnotes and images into a Folio document and pulls the bibliography into your library.
- New document → Import .docx.
- Drop in one file, or select several at once for a bulk import: one Folio document is created per file.
- Review what Folio found, then confirm.
Every reference is checked against CrossRef and OpenAlex. Only the ones Folio can verify are ticked by default; anything it can't confirm is flagged Unverified and left unchecked, so a mis-parsed or wrong citation never slips silently into your library. In a bulk import, the verified references from every file are merged into your library and de-duplicated across files.
Images come in as real figures. An image too large to carry over (over 1.2 MB, or past 3.5 MB of images in one file) is left as a placeholder you can replace.
Already use a reference manager?
If your library lives in Zotero, Mendeley, or a .bib/.ris file, use Add source → BibTeX / RIS, Zotero or Mendeley. These import cleanly, carry your existing metadata, and use no AI credits. If you save your Zotero API key under Account settings → Connected accounts, you can re-import later without pasting it again. See Importing and exporting your library.
What about email attachments?
Direct email connectors aren't available yet. For now, download those PDFs and bulk-import them. It takes seconds and keeps everything on your machine until you choose to upload it.
Frequently asked
Will importing overwrite anything? No. Imports only add, and duplicates are skipped.
Are my PDFs private? Yes. Uploaded PDFs are stored in a private bucket and served only to you. See How Folio stores your PDFs.
Can I do this later? Always. Every importer stays in New document and in Add source in the Sources library.