Where Rayyan stops
Rayyan is a popular screening tool with a workable free tier, and for title and abstract voting it does the job. As of 2026 the free plan caps you at three active reviews and puts the PRISMA flow diagram and duplicate auto-resolve on paid plans. And then the review continues and Rayyan does not: you export your includes and rebuild the project somewhere else for extraction, risk of bias, analysis, and reporting.
The screening itself
- Import RIS, or search PubMed and Europe PMC from inside the app and import results directly
- Deduplication on DOI and PMID, with every duplicate pair shown for review before it is dropped
- Blinded dual screening: reviewers vote independently, conflicts go to a resolver, agreement reports as Cohen's kappa
- AI relevance scoring against your inclusion criteria, with a recommendation and reason per reference; the decisions stay yours
- A separate full-text stage with retrieval status and exclusion reasons
- A PRISMA 2020 flow diagram that fills itself from your screening decisions
What happens after screening
The same project keeps going: data extraction with custom templates and AI drafts from your PDFs, risk of bias with RoB 2, ROBINS-I, NOS, JBI, and AXIS, meta-analysis and network meta-analysis validated against R, GRADE certainty ratings, and a manuscript draft assembled from what you actually did.
An honest note
Rayyan is free for basic use, its mobile app (full access on paid plans) is convenient for screening in spare minutes, and many large teams already run on it. If screening is the only step you need help with, it is a fine choice. The case for Covexe is the rest of the review living in the same place as the screening, with the PRISMA diagram in the free core rather than behind a plan.