Calibrated AI auto-excludes the clear nos — with an audit trail librarians can defend. You choose how deep (Replay · Plus · Monitor); screening runs at Careful, our conservative default. New accounts get 2 free title/abstract evaluations.
Using Covidence? AI pre-screening that cuts your screening pile before it reaches Covidence.
See how →Silent, captioned walkthrough using test reviews built from public PubMed records.
Using Covidence?
Keep the workflow your team knows. SyRev removes the clear exclusions against your own eligibility criteria, and the rest goes on to Covidence for normal dual screening.
Replay a review you already finished to see how much would have been cleared first.
Your methods paragraph, decision log and validation pack (a blind sample of the exclusions with a Wilson confidence interval) stay in SyRev for your write-up.
SyRev exports RIS files you can import into Covidence. SyRev is not affiliated with or endorsed by Covidence. Covidence is a trademark of its owner.
SyRev sells workload removed — with proof you can report.
2 free ti/ab evals · Replay $149$499 · Plus $249$899 · seats
SyRev Plus adds a full-text pass (open-access PDFs fetched when records have a DOI, PMID or PMCID; you upload the rest) — backed by four published findings on why abstracts alone are incomplete. See 1–4 on pricing →
A side-by-side of SyRev and three widely used review tools on price, AI screening, validation evidence, PRISMA, full-text screening, dual review, deduplication and data extraction, with every competitor detail sourced from their public websites (as of 25 September 2026).
Screening you can defend in your methods section, priced for a single review.
At our default setting, SyRev kept 96.5% of relevant studies (9,363 of 9,707) across 310 public review datasets, and auto-excluded 74–79% of records in full search pools from published systematic reviews. Every dataset and method is on our validation page. Then evaluate a finished search of your own: auto-exclude rate, recall check, hours saved and an audit appendix.
SyRev works from your review's own eligibility criteria, so your team doesn't label a single record before clear exclusions leave the queue. Reviewers spend time where judgment matters.
Dual review with AI as second reviewer comes with every plan, from a single $149$499 review. Your two free evaluations include it too.
Decision logs, conflict history, a ready-to-edit methods paragraph and a PRISMA 2020 / RAISE-style AI-use statement, ready for your manuscript and your peer reviewers. Per-criterion probabilities and a validation pack come with it.
Name the two reviewers for your records, or split them at random across a pool of up to 30 members, with the split recorded for your methods. Available on Team and institutional plans.
Replay is $149$499 and Plus is $249$899 per review, with no seat commitment. Team and institutional plans are there when your group grows.
Manual extraction is free on every plan, including your free evaluations. With Plus, AI pre-fill and an AI second extractor (with a compare-and-resolve step) cover up to 500 studies per review, and every suggested answer carries an exact quote and page number for you to check. Export CSV, Excel, a blank form or a RevMan Web pack, with a change history and methods text. It’s newly launched, so tell us what you think.
At Careful, 96.5% of relevant studies were kept across 310 public review datasets, with individual reviews sometimes lower. How much of the pile leaves depends on the review. Gold labels are often full-text or final-review decisions; Replay scores title/abstract only.
In full search pools from published systematic reviews (four benchmarks, 83 reviews), 74–79% of records were auto-excluded at Careful, while 94.9–97.4% of relevant studies were kept.
Environmental-health and preclinical sets (SWIFT-Review, 5 sets): 47.6% auto-excluded, with 99.7% of relevant studies kept.
Where records had already passed abstract screening (CSMeD-FT, 211 published systematic reviews), only 5.6% left; on a preclinical depression set (Bannach-Brown et al. 2019), 10.7%. When criteria barely show in abstracts, the pile barely shrinks — we show that on purpose.
Some individual reviews land well below the pooled figure — for example, 5 of 19 CLEF topics fell under 95% — often because of how the criteria were worded. For the datasets built from published systematic reviews, the criteria were generated automatically from each review's published selection criteria. None were changed after seeing results. Always prefer a run on your finished search over any vendor-style headline. See full results
One proof service. Pick the depth — then use your 2 free title/abstract evaluations on a review you already finished.
Academic is for universities, hospitals, non-profits and government. Commercial is for for-profit organisations and paid client work. Academic plans cannot be used for for-profit or paid client work. You choose at checkout. Full pricing
Title / abstract depth
Deeper depth · full-text
Conservative screening
Sign up, upload a finished search + criteria, and see auto-exclude rate, missed includes if any, hours saved, plus a methods paragraph and audit appendix. Use the second free evaluation on another finished review — or re-run with refined criteria. Paid Plus adds an AI full-text pass: open-access PDFs are fetched automatically (Unpaywall, Europe PMC) when records have a DOI, PMID or PMCID; you upload the rest.