Changelog
Releases
EvidenceFlow is under active development. We ship updates as features are built and verified rather than on a fixed schedule, and this page tracks every notable release — new capabilities, workflow changes, and the reasoning behind them — so teams running an active review always know what changed and why.
Versions follow semantic versioning: a major version (1.0, 2.0) marks a significant shift in workflow or data model, a minor version adds new functionality without breaking existing projects, and a patch release fixes bugs. Because projects, screening decisions, and extracted data all live in your own deployment's database, upgrading never requires re-entering data — existing projects carry forward automatically.
Initial release
- Literature import and deduplication
- Import citation lists from PubMed, RIS, BibTeX, and CSV. Records are automatically matched and deduplicated by DOI, PMID, and title/author similarity so reviewers don't screen the same study twice.
- Two-stage AI-assisted screening
- Title/abstract screening and full-text screening run as separate stages, each with AI-suggested include/exclude decisions and a confidence score reviewers can accept, override, or discuss.
- Full-text PDF retrieval
- Automatic full-text lookup across 10+ open-access sources, including PubMed Central, Unpaywall, Europe PMC, Semantic Scholar, OpenAlex, Crossref, bioRxiv/medRxiv, and CORE.
- Configurable data extraction
- Build a custom extraction form per project, then let AI pre-fill values from the full text in batch — reviewers verify and correct rather than typing every field by hand.
- Meta-analysis engine
- Fixed and random-effects pooling (inverse-variance, Mantel-Haenszel, Peto), subgroup analysis, heterogeneity statistics (Q, I², τ²), and interactive forest and funnel plots.
- Omics analysis
- GWAS and Mendelian randomization detection, with gene × study × variant visualizations for evidence synthesis across genomics literature.
- PRISMA flow diagrams
- A PRISMA 2020 flow diagram is generated automatically from your screening counts and updates in real time as records move through each stage.
- Report export
- Export results and reports as PDF, Word, Excel, CSV, or JSON, so review data stays portable outside the platform.
What's next
Upcoming work continues the RevMan-parity roadmap for meta-analysis — additional pooling methods, richer risk-of-bias summaries, and deeper subgroup reporting — alongside expanded free statistics tools and SEO-facing documentation. Nothing here is a firm commitment date; features ship when they're verified against known statistical references, not on a countdown.
Feedback
If a release changes something your review workflow depends on, or you'd like to request a feature, reach out through the contact page. Real usage from active systematic reviews and meta-analyses is what shapes this roadmap.