Orviva

What's new in Orviva

Release notes for the Orviva health-data platform. Orviva ships updates every week; all features are free during Early Access.

v5.3 — Methylation / Epigenetic Analysis

· milestone

  • New "methylation" top-level data category — CSV of CpG β-values (Illumina 450k / EPIC or WeGene methylation exports) is parsed by a pure Node.js math pipeline, no LLM
  • 5 epigenetic clocks (Horvath, Hannum, PhenoAge, GrimAge, DunedinPACE) computed from intercept + Σ wᵢ·βᵢ with coverage-gated confidence
  • 3 age accelerations (Horvath / PhenoAge / GrimAge vs chronological age from your profile) with ±3y reference bands
  • 2 methylation risk scores: AHRR smoking-exposure MRS and Epi-CRP chronic-inflammation predictor
  • 6-cell immune composition (CD4⁺ T, CD8⁺ T, NK, B, monocytes, granulocytes) via Houseman reference NNLS deconvolution
  • Per-sample QC: β-mean, detection-p fail rate, predicted sex from an X-chromosome sex-dimorphic panel
  • Dashboard: new "Biological Age / Epigenetic" trend group with English + 中文 locales so repeat methylation samples build a longevity trajectory
  • HealthDataDetail page renders 6 purpose-built cards per methylation document: biological age, DunedinPACE gauge, MRS scores, cell composition, age-acceleration bars, and QC
  • IndicatorDetail bands replace flat reference ranges for DunedinPACE, age accelerations, smoking MRS, and Epi-CRP with 4-color severity strips
  • METHYLATION_LAB_MAP cross-links epigenetic indicators to lab markers (CRP, APOE, TC, HBA1C, WBC) for cross-domain interpretation in the agent API
  • /health-data/profile now returns an `epigenetic` block with the latest snapshot (clocks, accelerations, MRS, cell proportions, QC)
  • G2 chat accepts methylation .csv attachments — results are parsed, written to the health vault, and summarised inline
  • Upload dialog auto-detects methylation via filename hints (methylation/cpg/bisulfite/epic/450k/850k/beta/甲基化) and a first-line CSV header sniff (cg-prefixed probe IDs vs rs-prefixed SNP IDs)

v5.2 — Medical Imaging Auto-Extraction & Cross-Page Description Matching

· feature

  • Auto-detect imaging pages (ultrasound, CT, X-ray, ECG, bone density, fundus) in checkup report PDFs
  • Each imaging exam is extracted as a separate document with the original page image uploaded to S3
  • Cross-page description matching: LLM matches imaging pages with their text descriptions from anywhere in the report
  • Auto-created imaging documents are linked to parent checkup report with 'Auto-extracted' badge
  • Cascade deletion: deleting a parent report also removes all child imaging documents and S3 assets
  • Health Profile API now includes imaging findings in the unified summary endpoint

v5.1 — CGM Time-Series Data, Dashboard Cross-Reference & Family Profiles

· milestone

  • CGM (Continuous Glucose Monitoring) support: upload Sibionics xlsx files for structured glucose analytics
  • Time-series framework: streams, data points, daily summaries, and period summaries with TIR, GMI, CV metrics
  • Dashboard CGM panel: average glucose, Time in Range, GMI, and 10-day trend chart
  • Cross-reference analysis: correlate genomic risk factors (e.g., Type 2 Diabetes) with lab results and CGM data
  • Unified upload entry: auto-detect CGM files vs genomic files vs health documents from a single upload button
  • Time-series data integrated into Health Data page as a new 'Time Series' tab
  • Health Profile API extended with timeSeriesData field for AI agent access
  • Family health profiles: manage multiple family members under one account with per-person data isolation
  • Cascading delete for health documents: records, findings, jobs, S3 files, and API keys all cleaned up
  • Health Profile summary mode: compact conclusions-only JSON (~2-4KB) for efficient AI agent conversations

v5.0 — Knowledge Base v5.0 — 122K+ SNPs, Re-analyze & Profile-Scoped Keys

· milestone

  • Upgraded to Gene2AI Knowledge Base v5.0-gwas-cpic: 122,896 SNP markers across 6,491 genes in 11 categories (131x increase from v3.5)
  • Re-analyze feature: re-submit genomic files to get fresh analysis with the latest KB — single job or batch re-analyze all
  • API Keys are now profile-scoped: each key is bound to a specific health profile for multi-family data isolation
  • Export filenames now include profile name for easy identification
  • Batch insert chunking: large genomic imports (17K+ records) now reliably complete without MySQL limits
  • Guide page: new Re-analyze vs Refresh documentation section
  • SEO metadata updated to reflect v5.0 knowledge base stats

v4.3 — API v3.5 Sync — 9 Categories, Dashboard Overhaul & Refresh

· feature

  • Synced with Gene2AI API v3.5: 933 SNP markers, 273 genes, 363 LD proxies across 9 analysis categories
  • 4 new genomic categories: APOE genotyping (ε2/ε3/ε4), HLA allele typing (9 alleles), CYP450 metabolizer phenotyping (CPIC), NAT2 acetylator typing
  • 5-tier health risk levels: low, average, slightly_elevated, elevated, high
  • Drug response now includes PharmGKB + CPIC data (192 markers, up from 162)
  • Dashboard overhaul: health risk overview panel, genomic risk summary, cross-reference insights (genomic ↔ lab data)
  • Genomic record cards redesigned: structured display with risk color-coding, phenotype badges, expandable details
  • HealthDataDetail: risk summary panel with elevated/low filter, groups sorted by risk priority
  • Refresh Analysis: re-fetch latest results from api.gene2.ai (single job or batch refresh all)
  • Direct .zip file upload support (no auto-compression needed)
  • Mobile UX improvements: larger touch targets, stacked layouts, better footer spacing

v4.2 — Direct Upload & Gene2AI API Migration

· improvement

  • Migrated all genomic analysis to Gene2AI's unified API (api.gene2.ai) — removed legacy WeGene API dependency
  • Frontend now uploads files directly to Gene2AI via presigned URLs, bypassing gateway size limits
  • Supports large genomic files via ZIP compression (up to 100MB) — compress your raw data before uploading for best results
  • Three-step upload flow: prepare → direct upload → confirm for maximum reliability

v4.1 — WeGene (微基因) Support

· feature

  • Added WeGene as a supported genomic data source alongside 23andMe and AncestryDNA
  • Auto-detection of WeGene files by filename pattern
  • WeGene data format passed through to analysis engine for optimized parsing
  • Updated all upload interfaces (Upload page, Upload dialog) with WeGene option

v4.0 — Trend Charts & Mobile Optimization

· feature

  • Interactive trend charts for numeric health indicators (line chart with reference range shading)
  • Trend overview on Health Data Vault page — prioritizes abnormal indicators
  • Optimized Health Data detail page for mobile (stacked layout, collapsible doc preview, compact spacing)
  • Backend API for indicator trend data (GET /api/v1/health/trends)

v3.5 — Unified Health Data Architecture

· improvement

  • Unified upload: accept genomic files (.txt/.csv) and health documents (PDF/images) from a single Upload button
  • Auto-detect file type — genomic files route to analysis, health documents to AI parsing
  • Unified My Jobs page: merged genomic analysis tasks and document parsing tasks in one view
  • Unified API Keys: single user-scoped key for all health data (genomic + medical + self-reported)
  • Streamlined navigation and entry points across the app

v3.2 — Health Data Vault — Agent API & OpenClaw Skill

· feature

  • Agent file upload API (POST /api/v1/health-data/upload) — upload documents from chat agents
  • Agent structured data submission (POST /api/v1/health-data/records) — record metrics via API
  • Agent document status query (GET /api/v1/health-data/doc/:id) — poll parse progress
  • OpenClaw Skill v2.0 — upload health docs, record daily metrics, query health data via Telegram/Feishu
  • Source tracking for multi-channel data collection (web, openclaw, telegram, etc.)

v3.1 — Health Data Vault

· milestone

  • New Health Data Vault page (/health-data) — centralized view of all health documents and records
  • AI-powered document parsing: upload lab reports, checkup results, medical records (PDF/images) and get structured data
  • Manual health data entry with step wizard and date emphasis
  • Document detail page with split view: original document + extracted results table
  • Auto-import genomic analysis results into Health Data Vault on completion
  • User-scoped API keys with Bearer token authentication for AI agents
  • Agent Health Data API: full records, delta sync, and summary endpoints
  • Category filtering, search, and document management (edit/delete)

v3.0 — Major Expansion — 903 SNP Markers

· milestone

  • Expanded from 37 to 903 curated SNP markers across 269 genes
  • Health Risks: 90 conditions analyzed (365 markers) — cardiovascular, cancer, diabetes, neurological & more
  • Drug Response: 52 drug interactions (162 markers) — Warfarin, Clopidogrel, Statins, SSRIs & more
  • Traits: 43 genetic traits (158 markers) — caffeine metabolism, alcohol flush, muscle fiber type & more
  • Nutrition: 37 nutrient insights (166 markers) — Vitamin D, Folate, Omega-3, Iron, Calcium & more
  • Ancestry: population composition (52 markers) with regional breakdowns
  • Population-aware analysis across 5 populations: EUR, EAS, AFR, SAS, AMR
  • 221 LD proxy variants (R² ≥ 0.7) for maximum chip coverage
  • Multi-chip support: 23andMe V3/V4/V5, AncestryDNA V1/V2, WeGene
  • LLM-enriched descriptions with confidence scores and population-specific notes

v0.2.1 — Early Access Launch

· milestone

  • Launched Early Access — all future updates included free
  • 37+ curated SNP markers across 5 categories: health risks, drug response, traits, nutrition, ancestry
  • AI-Ready JSON output with structured, machine-readable format
  • API key system for AI agent integration (ChatGPT, Claude, OpenClaw, custom bots)
  • Support for 23andMe and AncestryDNA raw data files (.txt, .csv, .zip)
  • LLM-enriched descriptions with clinical significance and confidence scores

v0.2.0 — Migrated to Self-Built Analysis Engine

· improvement

  • Switched to Gene2AI's own analysis API (api.gene2.ai) — all genomic data now processed through our unified pipeline
  • Improved analysis pipeline with better SNP matching and genotype interpretation
  • Added comprehensive error handling and retry logic for analysis jobs