Systematic reviews, from protocol to PRISMA.
Plan, search, screen, appraise, synthesize and report in one audited workflow — with blinded review teams, AI that assists, and reviewers who decide.
Every stage, in order, with a gate between each.
A review can’t move forward until the stage before it is genuinely complete — so nothing gets skipped under deadline pressure.
Start rigorous. Stay reproducible.
- PICO and verbatim eligibility criteria, locked before the search and versioned with every amendment
- Register the protocol on OSF from inside the review
- AI drafts the search strategy, you refine it — MeSH terms are checked live against NLM
- PRESS peer review of the search strategy before it runs
- PubMed and Europe PMC searched directly; RIS and NBIB import for other databases
- Duplicates merged by DOI and title matching, with every merge kept on record
Independent eyes. No shortcuts.
- Dual, blinded screening of titles, abstracts and full texts — each reviewer’s calls stay hidden until the other has submitted theirs
- Conflicts surface automatically and go to an adjudicator or third reviewer
- AI screens alongside as a blinded advisor, recorded separately from human decisions
- AI-proposed extraction fields with provenance; reviewers accept or correct every one
- Risk of bias with RoB 2, ROBINS-I, Newcastle–Ottawa, QUADAS-2 or AMSTAR 2, domain by domain
From evidence to publication.
- Reviewer work converges into one reconciled dataset
- Fixed- and random-effects meta-analysis (REML or DerSimonian–Laird, optional HKSJ intervals) with heterogeneity
- Egger and Begg tests for small-study effects
- GRADE certainty of evidence, rated per outcome
- PRISMA 2020 flow diagram and inter-rater agreement (Cohen’s κ) generated for you
- Exports and a full audit trail — then re-run the locked protocol as a living review
Humans decide. AI assists.
Run a review solo or with a dual-reviewer team. Give everyone a role, and R-Sys enforces who can see what, and when.
A blinded advisor, not a decider
The AI’s screening call is recorded next to reviewers’ decisions — never in place of them.
Traceable suggestions
Every AI-proposed extraction carries its provenance and runs on the model fixed in the locked protocol.
Audited end to end
Decisions, duplicate merges and exports are written to the review’s audit log.
Run it yourself, or run it with us.
R-Sys Platform
For review teams who know the method and want a rigorous, audited place to run it.
- The full workflow, protocol to PRISMA
- Team roles, blinding and conflict resolution
- AI-drafted search, advisory screening and assisted extraction
- Meta-analysis, GRADE, exports and audit trail
R-Sys Guided Review
For residents, departments and institutions who want experienced researchers alongside them, start to finish.
- Everything in the Platform
- Protocol and search strategy design with you
- Guidance through screening, appraisal and synthesis
- Manuscript, journal selection and peer-review support