Systematic review & meta-analysis

Evidence synthesis, in one rigorous workspace.

Search nine databases, screen with blinded co-reviewers, extract structured data, assess risk of bias and GRADE certainty, run meta-analysis, and export publication-ready figures. No R, Python, or desktop install.

Start free, no card See the workflow
9 databases PRISMA 2020 300 DPI exports 0 code required
verflux.com / project / analysis / forest-plot Publication ready
Verflux forest plot showing a random-effects meta-analysis with study confidence intervals and pooled estimate.
Search + reference workflow
PubMedEmbaseScopusEurope PMCCrossRefOpenAlexGoogle ScholarClinicalTrials.govWHO ICTRPZoteroMendeleyEndNote PubMedEmbaseScopusEurope PMCCrossRefOpenAlexGoogle ScholarClinicalTrials.govWHO ICTRPZoteroMendeleyEndNote

Science first

Less software friction. Same scientific responsibility.

Verflux supports careful evidence synthesis. It does not replace protocol design, critical appraisal, statistical judgement, or the researcher behind the work.

Output is publication-ready: 300 DPI figures, PRISMA 2020 flow diagrams, and methods text written for a manuscript.

Methods stay visible

Key estimators, assumptions, and source references are surfaced so the analysis is easier to inspect, explain, and defend.

Decisions stay connected

Search, screening, extraction, bias assessment, GRADE, and analysis live in one project instead of being scattered across tabs and spreadsheets.

Accessible without being simplistic

No-code workflows lower the technical barrier for students while preserving the structure serious review teams expect.

The pipeline

One evidence trail, from first search to final figure.

Each stage carries context forward. Screening decisions inform counts, extracted arms feed analyses, and appraisal stays attached to the evidence, which cuts manual handoffs and keeps the review easier to audit.

01 / Start with the evidence, not the spreadsheet

Start with the evidence, not the spreadsheet

Search nine databases from one workspace, then sync Zotero or import RIS from Mendeley and EndNote. Cross-source duplicate matching keeps the library clean before screening begins.

02 / Title, abstract, full text

Title, abstract, full text

Bulk include, exclude, or maybe. Blinded dual-reviewer mode holds each vote back until both are in, then routes disagreements to a conflict queue with Cohen's κ.

03 / Arm-level, not study-level

Arm-level, not study-level

Intervention and control captured separately for continuous and dichotomous outcomes. Auto-fill pulls study metadata from a DOI or PMID. Unlimited outcomes per study.

04 / Five validated instruments

Five validated instruments

RoB 2, ROBINS-I, Newcastle-Ottawa, QUADAS-2, and AXIS. Domain judgements roll up into the traffic-light table journals ask for, with an overall column.

05 / Certainty per outcome

Certainty per outcome

Five downgrade and three upgrade domains, rated outcome by outcome. The summary-of-findings table builds itself from the ratings.

06 / Pool it, then stress it

Pool it, then stress it

Random-effects and fixed-effect models with subgroup analysis, leave-one-out sensitivity, meta-regression, and Egger's test. Heterogeneity reported as I², τ², Q, and H².

07 / Figures a journal will accept

Figures a journal will accept

Forest plot with weight-proportional markers and a prediction interval, funnel plot with pseudo-confidence region, and a network diagram for direct comparisons.

08 / PRISMA 2020 and methods text

PRISMA 2020 and methods text

The flow diagram is built from your own screening counts and stays editable. Methods wording and exports in CSV, JSON, RIS, DOCX, and HTML.

Reference manager sync screen with Zotero, Mendeley, and EndNote tabs and API key setup. Screening table showing imported studies with include, exclude, and full-text decision controls. Data extraction dialog with separate intervention and control arm fields and an auto-fill action. Risk of bias traffic-light table for 20 studies across five RoB 2 domains with an overall judgement column. GRADE summary of findings table listing outcomes, downgrade domains, certainty ratings, and importance. Analysis screen showing pooled mean difference, prediction interval, heterogeneity statistics, and leave-one-out sensitivity table. Forest plot export with per-study estimates, weights, pooled diamond, and heterogeneity footnotes. PRISMA 2020 flow diagram with identification, screening, and included phases.

Reference manager sync01 / 08

The statistical engine

Statistics should be inspectable, not mysterious.

Verflux names the methods behind the result and keeps core estimator references close to the output. The goal is not a black box. It is a workflow that makes the statistical choices easier to verify, teach, and report.

PRISMA 2020 reporting Cochrane Handbook GRADE Working Group
Random-effects between-study variance
DerSimonian & Laird 1986
Alternative τ² estimator
Paule & Mandel 1982
Standardised mean difference
Hedges 1981
Fixed-effect inverse variance
Borenstein 2009
Small-study asymmetry test
Egger 1997
Rank correlation test
Begg & Mazumdar 1994
Fail-safe N
Orwin 1983
Inter-rater agreement
Cohen 1960

Compare

A complete review workflow, not another point solution.

Verflux keeps search, screening, extraction, appraisal, analysis, and reporting in the same project. That means fewer hand-offs, fewer spreadsheet copies, and one traceable path from retrieved records to the final figure.

Capability
Verflux
RevMan
R (meta)
Rayyan
Stata
No software install required
✓
—
—
✓
—
No R / Python needed
✓
✓
—
✓
—
9-database search built in
✓
—
—
—
—
Screening + extraction + RoB
✓
✓
—
Partial
—
GRADE certainty assessment
✓
✓
—
—
—
Forest & funnel plots
✓
✓
✓
—
✓
PRISMA 2020 flow diagram
✓
—
—
—
—
Real-time collaboration
✓
—
—
✓
—
Begg’s + Egger’s + Orwin’s N
✓
—
✓
—
✓
Cancel anytime
✓
Free
Free
Free
$500+
Works on any device / OS
✓
Win / Mac
✓
✓
Win / Mac

Comparison reflects the public capabilities currently described by Verflux; third-party products and licensing can change over time.

AI extraction + human sign-off

AI does the first pass. You approve every number.

Verflux reads the full text attached to each record and drafts every field with the sentence it was taken from. You accept, edit or reject. Nothing reaches your analysis unapproved, and anything the model cannot find in the paper is flagged instead of filled in.

Attach

Every record carries its full text

Attach the PDF to the record it belongs to. Verflux looks for an open-access copy by DOI through Unpaywall, and you upload whatever it cannot find. One copy is stored per unique file however many records point at it, so a paper reused across projects is stored once.

Automate

The AI drafts every field from the paper

One click drafts the whole extraction: design, setting, sample, arms, outcomes. Each value arrives with the quote it was read from, and a paper reporting several outcomes produces one set of arm data per outcome instead of forcing you to pick one.

Verify

You approve each number before it counts

Every draft value is checked against the source text first. Anything that cannot be found there is marked for you rather than quietly entered. You accept, edit or reject each field with the supporting quote in front of you.

Nothing reaches your extraction table, your forest plot or your export until you accept it.

Full-text storage is available on Professional and Unlimited. AI drafting is included on Unlimited and available as an add-on on other plans.

Review step 2 of 14 fields
arm1_mean high
-1.24

“Mean change in HbA1c from baseline was -1.24 percentage points in the intervention arm”

AcceptReject
arm2_sd not in text
0.81

This number is not in the source text. Check it against the PDF before accepting.

AcceptReject

An illustration of the review screen. The second card shows what happens when a proposed number cannot be found in the paper.

Researcher feedback

Built for the people doing the work.

Researchers come to Verflux for speed, but the deeper value is continuity: fewer handoffs, fewer duplicate files, and a clearer path from screening decisions to the final analysis.

“I completed a 61-study meta-analysis in under a week. What used to take months of juggling PubMed, Excel, and R now happens in one browser tab. The forest plots are publication-ready.”
SA
Dr. S. AhmedAssistant Professor, Internal Medicine
“The GRADE assessment and PRISMA 2020 flow diagram generation saved me enormous time. The statistical output matches R metafor, but I do not need to write a single line of code.”
MK
Dr. M. KhanSystematic Review Researcher
“Finally a tool that handles dual-reviewer screening AND meta-analysis in one place. The conflict resolution queue is brilliant. Highly recommend for any research team.”

Simple pricing

Monthly or annual. Cancel anytime.

Start with the full analytical workflow free for 3 days. Choose monthly or annual billing when you are ready, and add AI-assisted extraction to any paid plan.

Trial

$0

3 days · no credit card

1 active project0 collaborators
Start free trial

Included

  • Demo data import
  • Test all analytical tools
  • 9-database literature search
  • Screening & extraction
  • All risk-of-bias tools + GRADE
  • Meta-analysis + PRISMA 2020
  • Plots & documents (view only)
  • Exports locked until upgrade

Starter

$3 /mo

billed monthly · cancel anytime

10 active projects5 collaborators / project
Get Starter

Everything you need to publish

  • All analytical tools
  • 9-database literature search
  • Screening & extraction
  • 5 risk-of-bias tools + GRADE
  • Advanced meta-analysis
  • Forest, funnel & network plots
  • PRISMA 2020 diagram
  • CSV, JSON, RIS, HTML & Word export
  • AI-assisted extraction add-on

Professional

$5 /mo

billed monthly · cancel anytime

20 active projects10 collaborators / project
Get Professional

Everything in Starter, with more room

  • All analytical tools
  • 9-database literature search
  • Screening & extraction
  • All risk-of-bias tools + GRADE
  • Advanced meta-analysis
  • Forest, funnel & network plots
  • PRISMA 2020 diagram
  • CSV, JSON, RIS, HTML & Word export
  • Full-text PDF storage (2 GB)
  • AI-assisted extraction add-on

Unlimited

$50 /mo

billed monthly · cancel anytime

∞ active projects20 collaborators / project
Get Unlimited

The full platform, without project limits

  • All analytical tools
  • 9-database literature search
  • Screening & extraction
  • All risk-of-bias tools + GRADE
  • Advanced meta-analysis
  • Forest, funnel & network plots
  • PRISMA 2020 diagram
  • CSV, JSON, RIS, HTML & Word export
  • Full-text PDF storage (5 GB)
  • AI-assisted extraction included free
Monthly or annual. Cancel anytime. Free trial. No card required. Your work remains accessible. Upgrade when you need exports.

Academic use

Make the software part of the record.

When Verflux supports your review, cite the software and describe the methods used. Clear reporting makes the analysis easier for readers, reviewers, and future researchers to understand.

Rehman, N. U., Saif Ullah, K., & Tufail, U. (2026). Verflux: A browser-based platform for end-to-end systematic reviews and meta-analysis (Version 1.0) [Software]. https://verflux.com
Rehman NU, Saif Ullah K, Tufail U. Verflux: A browser-based platform for end-to-end systematic reviews and meta-analysis [Software]. Version 1.0. 2026. Available from: https://verflux.com
Rehman, N.U., Saif Ullah, K. and Tufail, U. (2026) Verflux: A browser-based platform for end-to-end systematic reviews and meta-analysis, Version 1.0 [Software]. Available at: https://verflux.com
@software{verflux2026, author = {Rehman, Naeem Ur and Saif Ullah, Khalid and Tufail, Usama}, title = {{Verflux}: A browser-based platform for end-to-end systematic reviews and meta-analysis}, year = {2026}, version = {1.0}, url = {https://verflux.com}, note = {[Software]} }
Literature screening, data extraction, risk of bias assessment (RoB 2, ROBINS-I, Newcastle-Ottawa Scale, QUADAS-2), GRADE certainty of evidence ratings, random-effects meta-analysis (DerSimonian-Laird and Paule-Mandel estimators), and forest and funnel plot generation were performed using Verflux (Rehman, 2026), a browser-based systematic review and meta-analysis platform.

Questions

Got questions?

Does Verflux replace methodological judgement or supervision?
No. Verflux reduces software friction and keeps the workflow structured, but protocol design, eligibility criteria, interpretation, critical appraisal, and final scientific judgement remain the responsibility of the research team. Students should still follow their supervisor, protocol, journal, and institutional requirements.
Do I need R or Python?
No. Every model runs server-side and the figures render in the browser. There is nothing to install and no environment to maintain.
Which databases does the search cover?
Verflux searches nine databases simultaneously with automatic deduplication. The current workflow includes PubMed, Embase, Scopus, Europe PMC, OpenAlex, CrossRef, Google Scholar, ClinicalTrials.gov, and the WHO ICTRP, and you can also sync Zotero or import RIS from Mendeley and EndNote.
How far can I get on the free trial?
All the way to a pooled estimate and an on-screen forest plot. The trial runs 3 days with one project and every analytical tool unlocked. Export is what a paid plan adds.
Is it a subscription?
No. Each paid plan is a single payment and the plan does not expire. You can add project slots later from the dashboard.
How do I know the statistics are right?
Each estimator is implemented from its primary reference and named in the output: DerSimonian-Laird and Paule-Mandel for between-study variance, Hedges for standardised mean difference, inverse variance for the fixed-effect model, Egger and Begg for small-study asymmetry, Orwin for fail-safe N. The analysis panel reports which estimator produced the result.
Which risk of bias instruments are included?
Cochrane RoB 2 for randomised trials, ROBINS-I for non-randomised studies, Newcastle-Ottawa for observational designs, QUADAS-2 for diagnostic accuracy, and AXIS for cross-sectional studies.
Can two reviewers screen independently?
Yes. Blinded dual-reviewer mode hides each vote until both are recorded, then sends disagreements to a conflict queue. Cohen's κ is calculated for the pair.
What happens to my data if I never upgrade?
It stays in your account and stays accessible. Export is the only thing that locks when the trial ends.
How do I pay from Pakistan?
Easypaisa or JazzCash, with PKR amounts shown at checkout. Stripe and PayPal are also available.
What happens if I don't upgrade?
Your account and all data remain permanently accessible. After the 3-day trial, export features are locked until you upgrade.
Can I use Verflux for multiple reviews?
Yes — paid plans support multiple active projects. Professional includes 3 projects, Unlimited includes unlimited projects.

Support better evidence

Give the science your attention, not the software plumbing.

Create a free account and work through the complete evidence-synthesis workflow before deciding whether Verflux belongs in your research process.

Start free, no card See pricing