Full support for pairwise, prognostic, DTA, NMA, and scoping study types. Workflows, PRISMA variants, risk of bias tools, and meta-analysis methods now adapt automatically based on your study type selection.

New forest plot variants (classic, rainforest, drapery), funnel plots (contour-enhanced, sunset, trim-fill), GOSH plots, Baujat plots, and NMA-specific visualizations including league tables and rankograms. All exports at 600 DPI with 6 journal presets.

All workflow state is now persisted server-side with auto-save. No more lost progress on page refresh. Eliminated localStorage dependency for pipeline data, ensuring reliable state management across sessions.

Complete diagnostic test accuracy meta-analysis with bivariate models (mada) and network meta-analysis with netmeta. Includes SROC curves, coupled forest plots, rankograms, and league tables — all powered by R statistical engines.

Comprehensive security audit: input sanitization across all API endpoints, URL validation, storage quota enforcement via database triggers, and removal of unsafe serialization patterns.

Added new forest plots, funnel plots, and sensitivity analysis charts with improved interactivity and export options. All visualizations now support custom styling and can be exported in multiple formats.

Upgraded AI models for better PDF parsing and data extraction. Extraction accuracy improved by 15% for complex tables and multi-column layouts. Added support for extracting data from scanned documents.

Resolved an issue where workflow sessions were not properly restored after page refresh. Sessions now persist correctly across browser sessions, and users can resume work seamlessly.