Direct comparison
AI Systematic Review Screening Tools Compared
Nested Knowledge, ASReview, DistillerSR, and EPPI-Reviewer compared: pricing, AI screening method, extraction/synthesis support, and which fits your review.
Side-by-side comparison
| Dimension | Nested Knowledge | ASReview | DistillerSR | EPPI-Reviewer |
|---|---|---|---|---|
| What it is | Cloud platform (AutoLit workflow) covering the full review lifecycle — search, screening, extraction, appraisal, and synthesis — inside one shared project. | Free, open-source screening-prioritization tool only. Reorders title/abstract screening via active learning; does not handle extraction, appraisal, or synthesis. | Cloud-based, enterprise-grade review-management platform from Evidence Partners covering screening, extraction, and risk-of-bias workflows with a full audit trail. | Web-based, full-pipeline platform from UCL's EPPI-Centre covering screening through both quantitative and qualitative/mixed-methods synthesis. |
| Developer / owner | Nested Knowledge, Inc. (Minneapolis-based company) | Utrecht University's AI-aided Knowledge Discovery (AIKD) lab | Evidence Partners (Ottawa, Canada) | EPPI-Centre, UCL Social Research Institute, University College London |
| Access / pricing model | Commercial; institutional or project-based pricing — verify current terms directly with the vendor before budgeting. | Free — Apache 2.0 open-source license. No purchase price or subscription fee; self-hosted (local install, institutional server, or your own cloud instance). | Commercial, quote-based. Separate “Academic & Student” and “Corporate & Government” tiers; no public pricing published. | Subscription-fee, not-for-profit cost-recovery model administered directly by UCL's EPPI-Centre. Browser-only; trial access available. |
| AI screening approach | AI-assisted features across search, screening, and extraction, positioned by the vendor as aligned with Cochrane's RAISE principles for human-checked, documented AI use. | Active-learning text classifier (e.g. naive Bayes or SVM on title/abstract text) that re-ranks the unscreened queue after every reviewer decision. Underlying method peer-reviewed in Nature Machine Intelligence (2021). | AI Re-Rank reorders the screening queue by predicted relevance; AI Classifiers can auto-answer closed screening questions; Smart Evidence Extraction (SEE) is a generative-AI extraction feature with human-in-the-loop review. | Active-learning re-ranking for citation screening, plus pre-built ML classifiers flagging likely RCTs/systematic reviews/economic evaluations, and the option to train a custom classifier from a team's own decisions. |
| Data extraction & synthesis | Structured extraction and synthesis run inside the same AutoLit project as screening. | Not supported — screening-prioritization only. Teams pair it with a separate extraction/synthesis tool. | Structured, form-based extraction (including AI-assisted SEE); generates PRISMA flow diagrams and standard reports. No native meta-analysis engine. | Structured, form-based extraction plus both quantitative synthesis (meta-analysis, R-based) and qualitative/mixed-methods synthesis (framework and thematic synthesis) — broader synthesis support than the other three. |
| Living / continuously updated reviews | Yes — Bibliomine and a living-review update mechanism for re-running searches and refreshing synthesis on a schedule. | No native living-review feature; screening prioritization only. | Supports continuous/living-review use through configurable, repeatable workflows as new literature is published. | Yes — an ML-driven update-subscription feature, trained on already-included studies, flags new potentially relevant literature for living systematic reviews and living evidence maps. |
| Regulated / audit-trail use | Positioned for HTA and pharma/medical-device evidence submissions; documents AI use for reviewer verification. | No built-in audit/validation layer. Screening rationale (stopping rule, seed set) must be reported by the team itself, per PRISMA methods reporting. | Full audit trail logging every screening and extraction decision; built for regulated pharma, medical-device, and HTA/HEOR use. | Structured, form-based decisions with appraisal built in; used heavily in public-policy and public-sector review work rather than marketed specifically around pharma regulatory audit trails. |
| Best for | Teams wanting an AI-forward, full-lifecycle platform for HTA or pharma/device evidence submissions, or ongoing living reviews. | Teams that already have a review-management workflow and want a free, transparent way to cut screening time on a large record set. | Enterprise, pharma, medical-device, and HTA/HEOR teams needing a validated, auditable, organization-wide review platform. | Teams wanting a not-for-profit, methodologically broad platform with native qualitative/mixed-methods synthesis alongside meta-analysis. |
Common questions
FAQ
Is ASReview really free to use?+
Yes. ASReview LAB is distributed on GitHub under an Apache 2.0 open-source license, so there is no purchase price or subscription fee for the software itself. Because it's self-hosted rather than a managed SaaS product, the only real cost is whatever compute you run it on — a local machine, an institutional server, or your own cloud instance.
Do any of these four tools eliminate the need for dual independent screening?+
No. PRISMA 2020 and the Cochrane Handbook expect two or more reviewers to independently assess each record, with conflicts resolved by discussion or a third reviewer. All four tools operate within that requirement rather than replacing it: ASReview and the active-learning features in Nested Knowledge, DistillerSR, and EPPI-Reviewer change the order records are presented to a human reviewer, or add a second AI check layer — they don't remove the human decision or the dual-review requirement itself.
Which of these tools support data extraction and synthesis, not just screening?+
Nested Knowledge, DistillerSR, and EPPI-Reviewer all cover extraction alongside screening. EPPI-Reviewer is the only one of the four with native support for both quantitative synthesis (meta-analysis) and qualitative/mixed-methods synthesis (e.g. framework and thematic synthesis) in the same platform. ASReview covers screening prioritization only — teams typically pair it with a separate extraction and synthesis tool.
Which tool is best for a living systematic review?+
Nested Knowledge (via its Bibliomine and living-review update mechanism) and EPPI-Reviewer (via its ML-driven update-subscription feature) both have purpose-built living-review support. DistillerSR can support continuous, repeatable re-screening as new literature appears, but it isn't built around a dedicated automatic-update mechanism the way the other two are. ASReview has no native living-review feature.
Can I use more than one of these tools in the same review?+
Yes, and it's common. Because ASReview is a screening-only tool, teams sometimes use it (or a similar active-learning classifier) to prioritize a very large record set, then move screened-in records into a full-pipeline platform like DistillerSR, Nested Knowledge, or EPPI-Reviewer for extraction, appraisal, and synthesis.
