Troubleshooting common issues

Import problems

"My CSV failed to import"

• Check that your CSV has columns for Title and (at minimum) Year. Authors, Abstract, DOI, and Journal are recommended.
• If exported from Excel, save as "CSV UTF-8" (not regular CSV) to handle special characters
• Maximum file size is 100MB. For larger files, split them into 2-3 smaller CSVs.

"My RIS file imported but records are missing fields"

• Some databases export incomplete RIS. Check that your search results include all required fields before exporting.
• Common missing field: Abstract. PubMed sometimes omits abstracts for older records.

"Duplicate detection missed some duplicates"

• LitRev matches duplicates on DOI, PMID, and title similarity (Jaccard > 0.85)
• Records with no DOI/PMID and substantially different titles won't be flagged as duplicates
• Manually mark duplicates from the project's Records page if needed

AI screening problems

"AI is screening but it's slow"

• Each record takes 8-15 seconds with both models. For 1000 records, expect ~3-4 hours
• Screening runs in the background — you can close the tab and come back
• Progress is shown in real-time on the Screening page

"AI made a wrong decision"

• Click the record to view it
• Override the decision with the Include/Exclude buttons
• Your override is final; the AI decision is replaced

"Why is the AI marking everything UNCLEAR?"

• Usually means your inclusion criteria are too vague
• Open your project settings → Criteria → make sure each criterion is specific (e.g., "Population: adult patients aged 18+ undergoing breast reconstruction" rather than "breast surgery patients")
• Re-run AI screening after updating

PDF problems

"PDF won't upload"

• Maximum PDF size is 50MB. For larger PDFs, compress them with Adobe Acrobat or ilovepdf.com
• Encrypted/password-protected PDFs aren't supported. Remove the password before uploading.
• Scanned PDFs without OCR can't be screened by AI. Run OCR with Adobe Acrobat or ocrmypdf first.

"PDF text isn't showing in the AI's quotes"

• The AI reads up to 20,000 characters from the PDF (typically the first 8-12 pages)
• If your full-text decision relies on content past page 12, you may need to upload a trimmed PDF containing just the relevant sections, or override the AI's decision based on your reading

Export problems

"PRISMA PNG export looks blurry"

• The PNG is exported at 2x resolution. Use it at the displayed size for crisp printing.
• For journals that require vector formats, use the PRISMA flow as a guide and recreate in your manuscript using a vector tool

"Methods text paragraph is generic"

• The auto-generated methods text uses your project's actual numbers but is a starting point
• Edit to match your manuscript's tone, add your database list, and customize the search dates

Still stuck?

Email support@litrev.app with:

• Your account email
• The project name where the issue happened
• A screenshot of the error
• What you were trying to do

We typically respond within one business day.