">
Back to Landing

The Platform

Genomic offset prediction.
0.015 milliseconds.

Three tools. One API. Enterprise-grade multi-tenant architecture with billing, audit logs, and sub-license management.

0.015ms
Per prediction on commodity hardware.
No GPU. No cloud required.
8/8
All independent scientific validation tests passed. Gradient monotonicity confirmed.
99.99%
Cost reduction vs field trial protocol. €280,000 → €0.000003 per result.

01 — Tool

Genomic Calculator

Prediction Engine
Available Now

Single and batch prediction engine. Accepts raw SNP dosage arrays and target environmental parameters — temperature, precipitation, pH. Returns genomic offset score (0–1), confidence interval, and planting recommendation in milliseconds.

Five-tier recommendation system: Highly Suitable / Suitable / Marginal / Risky / Avoid. Each tier calibrated against validated ecological thresholds.

CSV and VCF file upload — auto-mapped to SNP dosage format
Batch mode — up to 500 queries per API call with quota management
Prediction history — full audit trail per company
3D Adaptive Landscape — Three.js RDA projection surface
RDA scatter chart — training clusters with live query point
Environmental distance radar — standardised deviation display
🔬 Genomic Calculator

02 — Tool

ClimateZone Intelligence

🗺️ ClimateZone Intelligence
Geographic Analysis
Available Now

3D interactive WebGL map with genomic offset scoring across 12 Spanish climate zones. Real-time colour-coded suitability overlay. Markers pulse at their prediction frequency: green for suitable, amber for marginal, red for avoid.

Click any location marker for full environmental data — temperature, precipitation, pH, confidence score, and recommendation. Deployable to any geographic region.

4 ecotype preset buttons — switch ecotype, markers update instantly
45° pitch WebGL map with bearing offset for 3D depth
Persistent results bar — Suitable · Marginal · Avoid counts
Glassmorphism popups — full location data on click
MapLibre GL — offline-capable tile rendering

03 — Tool

Hex Intelligence Map

Regional Overview
Available Now

Regional suitability grid visualised as a hexagonal tessellation. Location name and offset score rendered inside each hex. Colour encodes genomic risk across micro-zones at a glance.

Hover any hex for a floating tooltip with full environmental data. Summary panel shows best and worst zones, Suitable/Marginal/Avoid counts, and active ecotype label.

Offset score rendered inside each hex — no legend lookup required
Canvas hover glow — shadowBlur effect on hovered hex
4 ecotype selectors — instant grid recalculation on switch
Summary panel — best zone, worst zone, counts
Canvas-native — no DOM overhead, scales to hundreds of hexes
Hex Intelligence Map

04 — Architecture

Enterprise Infrastructure

The platform is designed for multi-tenant deployment. Each licensee operates in an isolated company context with independent model training, prediction history, and quota management.

🔐
Authentication
JWT-based auth with PBKDF2-SHA256 password hashing. Token expiry, license expiry enforcement, 403 on expired access. Sub-company hierarchy with parent model inheritance.
💳
Billing Engine
Three billing modes: per-prediction (configurable price per call), subscription (fixed monthly), enterprise (manual contract, zero auto-charge). All transactions logged with royalty tracking.
📋
Audit Logging
Structured audit log on every critical action — login, training, prediction, billing change, admin action. Timestamped, user-attributed, queryable by event type. Root admin only.
🗄️
Model Persistence
Trained models serialised to SQLite and disk pickle. Startup priority: DB → disk file → fresh training. In-memory cache with lazy load fallback. Sub-companies inherit parent model.
Deployment
Single-command startup via bash run.sh. Auto port cleanup, dependency verification, .env generation, browser launch. Uvicorn with --reload for live development. Offline-capable.
🔌
API
FastAPI with OpenAPI docs at /api/docs. CORS open. Batch endpoint up to 500 queries. Full response models with Pydantic validation. Global 503 handler for DB unavailability.

05 — Competitive Reality

No existing tool delivers this.

We benchmarked every commercially available genomic offset tool. None provide real-time enterprise API access. None provide multi-environment batch queries. None are licensed for commercial deployment.

Feature R/vegan · gradientForest GenomicLand YTOE Platform ✦
Real-time API ✓ <0.02ms
Batch predictions Limited ✓ 500 / call
Multi-tenant architecture
Commercial licence ✓ Available now
Global climate map (50 regions)
Deployment time Weeks Weeks bash run.sh

* Performance benchmarks based on internal testing. R/vegan times measured on equivalent hardware.

06 — Integration

Two lines to a prediction.

The API is clean, documented, and requires no specialist knowledge to integrate. A standard HTTP POST returns a genomic offset score, confidence interval, and recommendation within milliseconds.

Request — POST /predict
{
  "genotype": [1.8, 0.2, 0.2, 1.8, 0.2],
  "environment": {
    "temperature": 18.4,
    "precipitation": 720,
    "soil_ph": 7.1,
    "date": "2026-09-01"
  }
}
Response — 200 OK · <15ms
{
  "genomic_offset": 0.1842,
  "confidence": 0.94,
  "recommendation": "Suitable",
  "predictions_remaining": 9847,
  "model_version": "v2.1.0"
}

The platform is operational and ready for licensing.

Enter Platform →