Skip to content
metaproc
Launch

Reproducible meta-analysis

Run a complete meta-analysis in your browser.

From raw data to forest plot to GRADE summary, with the exact R code behind every step.

Prefer the desktop app?
Launch free

A blank workspace to a defensible report, in four moves.

  1. Plan

    Frame your review first: the question, the studies you'll include, and the analyses you intend to run.

  2. Import

    Upload a CSV or Excel workbook, or load a bundled example. MetaProc detects headers and flags duplicate or out-of-range studies as you go.

  3. Analyze

    Choose a template, map your columns, and run. Forest and funnel plots, heterogeneity, subgroups, and sensitivity analyses appear instantly.

  4. Report

    Build a reproducible report with a GRADE summary, then export a bundle that re-runs in a clean R session and reproduces your pooled estimate.

MetaProc home screen: a welcome banner with Start: import your data and Plan my analysis buttons, above a five-step Your workflow strip (import data, explore (EDA), run an analysis, plan a pipeline, build a report).
The home screen mirrors those steps, and your whole project saves to a file you can reload later.

A workspace built around the code it runs.

Four corners of MetaProc. Each result carries the exact R that produced it.

Live R code panel: a metafor script with escalc effect-size, rma model, forest and funnel calls beside the raw random-effects model output it produced.
Live R code panel Every analysis shows the exact R it ran, metafor calls included. Copy it, audit it, re-run it.
MetaProc home screen with the Workflow panel open on the right, holding a saved risk-ratio analysis with buttons to send it to the Plots studio or the Report.
Workflow store Collect every result across tabs, then send it straight to Plots or Report.
Plots studio: a ten-study forest plot of risk ratios with a random-effects pooled estimate of 0.69, plot-type and drawing-engine controls on the left, and a Reproducible R panel underneath.
Plots studio Forest and funnel plots re-drawn from your saved analysis, code shown beneath.
Reproducible report builder with checkboxes for methods, PRISMA flow, included studies, pooled result, forest plot, heterogeneity, reproducible code and packages used, plus HTML, PDF and R Markdown output formats.
Report builder Export HTML, PDF or .Rmd, or a reproducibility bundle that re-runs anywhere.
See the full feature tour Watch the 73-second demo

Validated, not improvised

Methods you can defend

MetaProc drives the same gold-standard R engines the field already trusts, and it is golden-tested against them: 2,600+ automated tests check that every result matches the reference engine, byte for byte.

How we validate
  • metaforpooling and diagnostics
  • netmetanetwork meta-analysis
  • metareference dependency
  • robvisrisk-of-bias figures
  • esceffect-size conversions

Frequently asked

Is it free?

Free while in beta. We will give plenty of notice and a fair offer before anything changes.

Is my data private?

Analyses run in ephemeral, RAM-backed isolated containers hosted in the EU and destroyed at the end of your session. Work on the aggregate, study-level data a meta-analysis needs, not patient-level records.

Can I see the code my analysis ran?

Yes, that is the whole point. Every model, table, and figure is produced by building R code from your choices, running that exact code, and showing you the same text in a copyable panel. What you see is what ran.

More in the Quickstart

Your next meta-analysis starts in the browser.

No install, no setup. Free while in beta.

Launch MetaProc free