Every tool you need for a systematic review
ReviewBrewery automates the most time-consuming stages of evidence synthesis — screening, extraction, risk of bias, meta-analysis, and reporting — while keeping humans in the loop for every critical decision.
Get Early AccessMulti-Model AI Screening
Three independent AI models (GPT-4o, Claude Sonnet, Gemini Flash) screen every paper. A tiebreaker resolves conflicts. You review and override any decision.
- Screen 1,000+ abstracts in minutes, not months
- Multi-model ensemble eliminates single-model blind spots
- Full human-in-the-loop — accept, reject, or flag any AI decision
- Exclusion reason tracking with automatic PRISMA-ready categorization
Smart Deduplication
Automatically detect and remove duplicate records across databases using advanced fuzzy matching algorithms.
- Cross-database duplicate detection (PubMed, Scopus, WoS, Embase)
- Fuzzy matching handles title variations and encoding differences
- Review and restore any flagged duplicates
AI-Assisted Data Extraction
Define custom extraction schemas, then let AI extract study characteristics, outcomes, and population data from full-text PDFs.
- Custom extraction schemas — define exactly what data you need
- Full-text PDF import with automatic chunking and vector search
- AI extracts data points with source citations from the text
- Human verification for every extracted value
Risk of Bias Assessment
Assess risk of bias using multiple validated instruments with AI-assisted pre-population and full manual control.
- Cochrane ROB 2.0, Newcastle-Ottawa Scale, JBI checklists, and more
- AI pre-populates assessments as a starting point for human review
- Per-domain judgments with supporting justification text
- Visual summary tables for publication
Meta-Analysis Workbench
Run fixed-effect and random-effects meta-analyses with full statistical output — heterogeneity, prediction intervals, publication bias tests, and forest plots.
- Inverse-variance, Mantel-Haenszel, and Peto pooling methods
- Effect sizes: MD, SMD (Hedges' g), OR, RR, RD
- Heterogeneity (I², τ², Q), prediction intervals, Egger's test
- Subgroup analysis and sensitivity analysis
PRISMA Reporting
Automatically generate publication-ready PRISMA 2020 flow diagrams and reporting checklists from your screening decisions.
- PRISMA 2020 flow diagram with exclusion reasons at each stage
- Export as SVG or PNG for direct journal submission
- Exclusion reason merge and deduplication tools
Team Collaboration
Multiple reviewers screen, extract, and assess simultaneously with built-in conflict resolution and inter-rater reliability tracking.
- Real-time multi-reviewer screening with role-based access
- Conflict detection and resolution workflows
- Inter-rater reliability metrics (Cohen's kappa)
- Activity dashboards and progress tracking
AI-Generated Reports
Generate written report sections following the PRISMA 2020 checklist. AI drafts each section using your review data, and you edit and finalize.
- 24 sections following the PRISMA 2020 reporting checklist
- AI drafts with your actual data (screening stats, pooled estimates, etc.)
- Export to PDF or copy as Markdown
Flexible Import & Export
Import search results from any major database and export your data in standard formats.
- Import RIS, BibTeX, NBIB (PubMed), and EndNote XML
- Direct PubMed search integration
- Full-text PDF upload with vector-powered semantic search
- Export to CSV, PDF, and structured formats