What the Auto-Resolver does
The Auto-Resolver lets you define matching criteria and automatically resolve all duplicates that meet your rules in a single action. It is designed for larger datasets where resolving duplicates one by one would take too long.
Instead of reviewing each pair manually, you specify what makes two records a duplicate — exact field matches, a similarity threshold, or both — and Rayyan applies those rules across your entire dataset.
How to access and run the Auto-Resolver
- Go to the Review data tab and run duplicate detection first if you have not already done so
- Click Auto Resolve from the Review data tab or from within the manual resolution view
- Configure your matching criteria (see Criteria explained below)
- Review your settings carefully before confirming
- Click Auto Resolve to run
Note The Auto-Resolver is available on paid plans only.
Important Auto-resolution is irreversible. There is no undo — if you need to recover a deleted duplicate, see How to Recover a Deleted Duplicate in Rayyan. Always verify your settings carefully before confirming any Auto-Resolver run.
Criteria explained
| Criterion | Description |
|---|---|
| Exact match fields | Match records that are identical in one or more fields: Title, Author, Journal, Year, Pages, DOI, Publication Type. You can combine multiple fields for more precise matching (e.g., Title + Author + Publication Type). |
| Overall Similarity Percentage | Resolve duplicates that exceed a similarity threshold you set (e.g., 80% or 95%). |
| Imported references priority | Specify which imported dataset to treat as dominant when duplicates are found across different searches. |
| Text normalization | Converts text to lowercase and ignores punctuation for more consistent comparisons. Default: on. |
| Customization protection | Protects records that already have screening decisions from being automatically deleted. Default: on. |
Tips for best results
- Start with stricter criteria — begin with exact field matching or 100% overall similarity to resolve the most obvious duplicates first, then lower the threshold in a second pass to catch near-duplicates
- Use the Similarity Score during manual review — check the confidence scores on your unresolved pairs before deciding on your Auto-Resolver similarity threshold
- Keep Customization Protection on — this prevents records with existing screening decisions from being automatically deleted
- Check remaining unresolved count after running — some duplicates may not meet your criteria and will need manual handling
Continue to: How to Recover a Deleted Duplicate in Rayyan