Search, screen, appraise, and run a full meta-analysis in your browser: No R, Python, or installation required.
If you have ever tried to complete a meta-analysis, you know the work rarely lives in one place. You search in PubMed, deduplicate in a reference manager, screen in a separate app, extract into a spreadsheet, score risk of bias in a Word template, and then — if you can — open R or Stata to pool the results and draw a forest plot. Each handoff is a chance to mistype a number or lose track of a decision, and the final statistical step stops a lot of people cold.
Verflux is a browser-based platform that runs the entire systematic review and meta-analysis workflow in one place, with no programming and nothing to install. You search, screen, extract, appraise, analyse, and generate a PRISMA 2020 report from the same project, in any web browser, on any operating system. This guide explains how that works, what it can and cannot do, and how it compares to the tools you are probably using now.
A systematic review answers a focused question by finding every relevant study under a pre-defined, reproducible protocol, then appraising and summarising them. A meta-analysis is the statistical part: it combines the results of those studies into a single pooled estimate, with a measure of how much they disagree (heterogeneity).
Done properly, the process moves through a fixed set of stages: write a protocol, search the literature, screen records, extract data, assess risk of bias, rate the certainty of the evidence, run the analysis, and report it against the PRISMA 2020 standard. The methodology is well established. The friction is almost entirely in the tooling.
The statistical stage is the most common point of failure. The packages that dominate meta-analysis — meta and metafor in R, or commercial software like Stata — assume you can write code. For many clinicians, students, and researchers, that is a genuine barrier, and commercial licences are often unaffordable. Screening-focused web tools help with study selection but usually stop before meta-analysis, certainty rating, and standards-based reporting, so you end up exporting your data into yet another program anyway.
The result is a workflow stitched together from four or five tools that were never designed to talk to each other. Verflux was built to collapse that into one.
Everything in Verflux lives inside a project that holds your searches, screening decisions, extracted data, appraisals, analyses, and report files together, so the trail of how a result was produced stays intact. The workflow runs through nine stages.
Start with a structured protocol built on the PICO(S) framework, with a field for your PROSPERO registration number. Export it as a standalone document for registration or an appendix.
Run one search across roughly ten major sources at once, including PubMed/MEDLINE, Scopus, Embase, OpenAlex, CrossRef, Europe PMC, Web of Science, Google Scholar, the Cochrane CENTRAL register, and the ClinicalTrials.gov trial registry. Duplicates are detected and merged automatically. You can import references from RIS or BibTeX files, and a citation-chasing tool finds the papers that cite an included study (forward) or are cited by it (backward).
Screen at title/abstract and full-text stages with bulk actions and keyboard shortcuts. A blinded dual-reviewer mode records two independent decision sets and queues the disagreements for resolution, and inter-rater agreement is reported as Cohen's kappa.
Capture data at the study-arm level for continuous and dichotomous outcomes, with auto-fill from a DOI or PMID and no limit on the number of outcomes per study.
Use five validated instruments — RoB 2 for randomised trials, ROBINS-I for non-randomised studies, the Newcastle–Ottawa Scale, QUADAS-2 for diagnostic-accuracy studies, and AXIS for cross-sectional studies — feeding an automatic traffic-light summary.
Rate the certainty of evidence per outcome across the GRADE domains and produce a summary-of-findings table.
Run fixed-effect and random-effects meta-analysis on the standard effect measures (mean difference, standardised mean difference, risk ratio, odds ratio, risk difference), with subgroup analysis, meta-regression, cumulative meta-analysis, leave-one-out diagnostics, tests for small-study effects, and a frequentist network meta-analysis for comparing several treatments.
Generate forest plots with weight-proportional markers and a prediction interval, and funnel plots with pseudo-confidence regions, all exportable at 300 DPI.
Build a PRISMA 2020 flow diagram straight from your screening counts, generate a draft methods paragraph and a 27-item PRISMA checklist, and export your data as CSV, JSON, RIS, HTML, or Word.
Across all of this, projects can be shared with co-reviewers who screen independently, with a collaboration board and real-time chat alongside the data. A project can also sync two-way with a Zotero library, and records from Mendeley or EndNote import through RIS.
This is the question that brings most people here, so to be direct: yes, you can run a complete meta-analysis in Verflux without writing a single line of code. The statistical engine uses the same estimators as established packages — inverse-variance pooling, DerSimonian–Laird and Paule–Mandel between-study variance, Cochran's Q, I², H², and prediction intervals — and each formula follows its original published source. You enter your data and read the results; the computation happens on the server.
Verflux builds your PRISMA 2020 flow diagram from the record counts it already tracked during screening, which means the published figure and the underlying numbers cannot drift apart. There is no separate diagram tool to fill in by hand and no risk of the boxes disagreeing with your data.
Forest and funnel plots render in the browser and export at print resolution. For small-study effects you get Egger's regression, Begg's rank correlation, the trim-and-fill method, and Orwin's and Rosenthal's fail-safe N — the standard battery, without opening a statistics package.
When your question involves more than two treatments, Verflux runs a frequentist network meta-analysis that ranks the options with P-scores, presents every pairwise contrast in a league table, and draws a network diagram that separates direct from indirect evidence. It covers connected networks of two or more treatments; more specialised extensions such as component and dose–response models are on the roadmap.
| Capability | Verflux | RevMan | R | Rayyan | Stata |
|---|---|---|---|---|---|
| No software installation | ✅ | ❌ | ❌ | ✅ | ❌ |
| No coding required | ✅ | ✅ | ❌ | ✅ | ❌ |
| Multi-database search built in | ✅ | ❌ | ❌ | ❌ | ❌ |
| Forward/backward citation chasing | ✅ | ❌ | ❌ | ❌ | ❌ |
| Screening + extraction + risk of bias | ✅ | ✅ | ❌ | Partial | ❌ |
| GRADE certainty assessment | ✅ | ✅ | ❌ | ❌ | ❌ |
| Forest and funnel plots | ✅ | ✅ | ✅ | ❌ | ✅ |
| Egger's / Begg's / trim-and-fill | ✅ | ❌ | ✅ | ❌ | ✅ |
| Subgroup and meta-regression | ✅ | Partial | ✅ | ❌ | ✅ |
| Network meta-analysis | ✅ | ❌ | ✅ | ❌ | ✅ |
| PRISMA 2020 flow diagram | ✅ | ❌ | ❌ | ❌ | ❌ |
| Reference-manager sync (Zotero) | ✅ | ❌ | ❌ | ❌ | ❌ |
| Real-time collaboration | ✅ | ❌ | ❌ | ✅ | ❌ |
The pattern is straightforward: the R and Stata ecosystems are powerful but demand programming, RevMan and Rayyan each cover part of the workflow, and Verflux is built to cover the whole thing in one browser tab without code.
Verflux fits researchers who need a rigorous method but do not have the time, training, or budget for a code-based pipeline: graduate students running their first review, clinicians synthesising evidence alongside clinical work, research teams who want one shared project instead of a folder of spreadsheets, and educators teaching evidence synthesis who want students to follow the canonical steps rather than wrestle with software. Because nothing installs and the computation runs server-side, results do not depend on anyone's local setup — which matters when a review is produced by a team or used in a classroom.
You can try the full workflow before entering any of your own data: a demonstration dataset loads into a new account so you can walk through searching, screening, analysis, and reporting end to end. When you are ready, start your own project from the protocol stage and work straight through to the PRISMA report.
Start free at verflux.com
Can I do a meta-analysis without knowing R or Python? Yes. Verflux runs fixed-effect and random-effects meta-analysis, heterogeneity statistics, forest and funnel plots, and more entirely in the browser, with no coding. The underlying estimators match those used in standard statistical packages.
Is Verflux a free RevMan or Rayyan alternative? Verflux covers more of the workflow than either: where Rayyan focuses on screening and RevMan on appraisal and analysis, Verflux spans protocol, searching, screening, extraction, risk of bias, GRADE, meta-analysis, and PRISMA reporting in one tool. It offers a free trial; see the site for current plans.
Which databases can Verflux search? About ten major sources, including PubMed/MEDLINE, Scopus, Embase, OpenAlex, CrossRef, Europe PMC, Web of Science, Google Scholar, the Cochrane CENTRAL register, and ClinicalTrials.gov.
Can Verflux generate a PRISMA 2020 flow diagram? Yes. It builds the PRISMA 2020 flow diagram automatically from your screening counts and also generates a 27-item PRISMA checklist.
Does Verflux support network meta-analysis? Yes — a frequentist network meta-analysis with P-score treatment rankings, a league table of pairwise comparisons, and a network diagram. Component and dose–response models are planned for future releases.
Can I collaborate with co-reviewers? Yes. Projects can be shared, two reviewers can screen independently with automatic conflict detection and a Cohen's kappa agreement summary, and a collaboration board and real-time chat sit beside the data.
What can I export? Forest and funnel plots at 300 DPI, a PRISMA 2020 flow diagram and checklist, a draft methods paragraph, and your data in CSV, JSON, RIS, HTML, or Word.
Does it work on my computer? Verflux runs in any modern web browser on any operating system. There is nothing to install.
All analytical features included in the free trial. No credit card, no installation, no R or Python.
Create free account