AI literature screening: Rayyan vs DistillerSR vs ASReview

Turn a reference set into a review workflow

Systematic-review screening starts with a set of references and a defined inclusion question. Rayyan, DistillerSR, and ASReview LAB help reviewers find relevant studies, but the word AI covers different functions in each product. Reordering a queue, proposing an inclusion decision, and extracting study data are separate capabilities. The comparison below follows those differences and the workflow around the reviewer’s decisions.

Rayyan

Rayyan provides a shared environment for importing references, detecting duplicates, screening titles and abstracts, and continuing into full-text decisions and structured data. Its AI relevance predictions help reviewers prioritize attention while collaboration features keep inclusion choices, labels, and conflicts in the review workflow. Sampling offers a way to divide subsets among team members, and the mobile app supports screening away from the desktop. That breadth makes Rayyan useful for researchers who want a recognizable systematic-review workspace rather than a separate modeling tool. The plan boundary is especially important. Basic relevance predictions appear in the Free plan, while Advanced adds AI PICO extraction and larger review and reviewer allowances. ResearchPilot, with more extensive AI screening and extraction capabilities, is associated with institutional tiers. Rayyan suits academic and research teams that want collaboration, screening, and reporting in one service, with a package they can expand. Its basic predictions should not be described as equivalent to the institutional ResearchPilot feature set.

DistillerSR

DistillerSR combines configurable literature-review workflows with AI screening, evidence extraction, and reuse across an organization. Its AI Rerank learns from reviewers’ inclusion and exclusion decisions and reprioritizes references as work continues. AI error checking can flag potentially mistaken exclusions, while classifiers can help label and route references. The newer Smart Screening capability, launched in September 2026, evaluates titles and abstracts in a human-in-the-loop or automated workflow, with decision rationales and tracked activity. That is distinct from simply sorting a queue. Smart Evidence Extraction then supports finding and suggesting answers from source documents for extraction forms. DistillerSR suits evidence-generation teams that need screening to connect with configurable forms, reporting, and reusable evidence rather than stop at an exported shortlist. Its value depends on selecting the appropriate capabilities and human oversight for the work. The vendor’s processing capacity and pilot statistics do not establish the quality of a team’s own review or the allowances in its subscription.

ASReview LAB

ASReview LAB focuses on active learning for screening textual datasets. Reviewers import a reference set, provide relevant or irrelevant labels, and let the model prioritize which record to inspect next. The reviewer remains the source of inclusion decisions; the model learns from those decisions to make the queue more useful. The current open-source project offers configurable model components, automatic saving of labels, and exports of labeled data or project files. A simulation toolkit can assess model behavior on fully labeled datasets, giving researchers a way to study the screening method itself. Current version-three documentation also describes hiding duplicate records during screening and editable tags. ASReview LAB suits researchers who value transparent models, reproducible projects, and control over where the software runs. Its free software license comes with the responsibility for installing or hosting the application. It is a more focused screening companion than a managed evidence platform with procurement, extraction, and organization-wide administration bundled together.

Prioritization versus screening decisions

ASReview LAB centers on choosing the next useful record for a human reviewer. Rayyan includes relevance predictions in its basic workflow and reserves more extensive ResearchPilot functions for institutional scope. DistillerSR offers both adaptive reranking and a separate Smart Screening decision capability. These distinctions matter when comparing claims about AI screening. A team wanting to keep every inclusion decision with a reviewer can use prioritization; a team seeking proposed or automated decisions needs the relevant capability and the corresponding review workflow. Those are different purchases and different operating choices.

What happens after the shortlist

Rayyan continues through full-text screening, data extraction, and reporting within its review environment, with access varying by tier. DistillerSR places configurable extraction and evidence reuse alongside screening, making it relevant to organizations maintaining several reviews or recurring evidence programs. ASReview LAB saves labels and exports the dataset or project, but its central purpose remains screening prioritization. Rayyan is a natural fit for a collaborative review team, DistillerSR for a wider evidence-management process, and ASReview for a research group that wants the screening model to remain a distinct, controlled component.

Free software, free plans, and institutional AI

ASReview LAB has no recurring software license fee, although hosting and operation still take resources. Rayyan’s Free plan includes three active reviews and relevance predictions; Advanced is $8.33 per seat per month billed annually, while ResearchPilot belongs to institutional scope with separate minimums. DistillerSR’s current amount and contract interval require a proposal defining the AI modules. The useful cost comparison is therefore a self-managed screening tool, a hosted review package, or an evidence platform with specific automation capabilities. A free entry point does not imply that every AI function is included.

Which should you choose?

Consider Rayyan for collaborative reviews with a tiered path into AI features, DistillerSR for configurable screening connected with enterprise evidence management, and ASReview LAB for transparent active-learning prioritization you can operate yourself. Match the choice to the reviewer’s role and the work after screening.