Launching today
BeyondMeta
Meta-analysis you can defend — no coding, no install
4 followers
Meta-analysis you can defend — no coding, no install
4 followers
BeyondMeta is one workspace for evidence synthesis. Search eight literature sources, deduplicate, screen blind with your team, then run pairwise or network meta-analysis on deterministic R. GRADE, bias plots, PRISMA and a full reproducibility bundle included.
Hey Product Hunt — we're the team behind BeyondMeta. There are two of us: Shozab Mehdi and Shaheer Hyder.
A systematic review today runs across a pile of disconnected tools: one for the literature search, a reference manager for duplicates, a spreadsheet for screening decisions, then R or RevMan for the synthesis. The evidence trail breaks at every handoff — and by submission time, nobody can reconstruct how a given number was produced.
BeyondMeta puts the whole review in one workspace. Search eight free literature sources from one saved strategy, and import RIS, NBIB, BibTeX, PubMed XML or Web of Science exports from licensed databases. Deduplicate with likely copies side by side, keeping the canonical record and the full source trail. Screen blind, with two or more reviewers, calibration rounds and a separate conflict queue. Then synthesise pairwise or network on deterministic R — metafor and netmeta — with league tables, SUCRA and rankograms that always show the interval, never a bare winner list. Appraise and report with GRADE certainty, RoB 2 and ROBINS-I plots, and a PRISMA 2020 flow whose counts reconcile as you type. Export a reproducibility bundle: the R script, cleaned data, results and figures, packaged so a reviewer can rerun exactly what was run.
Two things we would not compromise on: automation proposes but you approve every scientific decision, and the statistics stay deterministic. Two GRADE domains never suggest anything at all, because they need the studies actually read — and our AI modules are marked in the lab rather than shipped as magic.
It runs in the browser. No coding, no installation. We are in open beta, and it is free while we are there — no card, no trial clock. At this stage we would rather have review teams actually using it and telling us where it breaks than charge anyone for it.
So if you have run a review before, we would really like to know: where did your evidence trail break? That is what we want to fix next.