One satellite scan · Many uses · EU-grade infrastructure

Satellite Intelligence
for Agricultural Land.

Crop monitoring, ragweed detection, subsidy validation, plant-health surveillance — all from one browser. Draw a field, run analysis, get results in 2 minutes. From €4/ha/year.

🚀 Start Free — No Card Required ▶ See the Platform
€4
Starting price · ha · year
9
Spectral indices calculated
10+
Crop types recognised
2min
Draw field · get results
🌾 Real fields. Real cultivation. Real intelligence.

From ground to orbit,
every day of the season.

The platform sees every cycle of your operation — soil preparation, planting, growth, irrigation, stress, harvest. Same satellite, same field, continuous monitoring. Where your crop is, where it's stressed, where the work is needed.

Aerial geometric farmland — green crop field, fallow soil, road, tree
10 m / pixel
Aerial Farmland
Every parcel, mapped.
10m resolution from orbit. Every field, every boundary, every crop type — automatically detected and continuously updated.
Aerial view — combine harvester and grain cart on a corn field
Harvest Detected
Operations Tracking
Every operation, timestamped.
Soil prep, planting, irrigation, harvest — every change in the field is detected and timestamped within days, not weeks.
Top-down aerial view — combine harvester with grain trailer in a corn field
Multi-sensor · revisit 3d
Satellite Monitoring
Sub-weekly revisit.
Radar + multispectral satellite imagery in one fusion. Cloud-resilient, every field re-scanned multiple times per week, automatically.
Top-down view of green crop field with visible machinery passes and tree line
NDVI · 0.82
AI Insights
Pixels into action.
BioGuardNet turns satellite pixels into decisions: which field, what crop, what stress, what stage, what to do next.
Live Platform · Real Data
🔒 app.agriguard.com /dashboard
AgriGuard Mission Control: country list, AreaScan widget, satellite map with field overlays, FieldGuard panel, stress alerts, NDVI timeline
Live · 9 countries · 407 boards
Mission Control · AreaScan · BioGuardNet inference
Real-time stress alerts · FieldGuard sync
Platform Features

Everything you need.
Nothing you don't.

Draw a field. Run analysis. Get satellite-grade intelligence in 2 minutes — no installation, no training required, no GIS team.

🛰️
Satellite Monitoring
Free EU satellite data processed automatically. 9 spectral indices: NDVI, NDRE, NDWI, EVI2, LAI, BSI, CRE, PSRI, SAVI — calculated per field on demand.
Live
🌿
Ragweed Detection
Automatic ragweed risk scoring. Early detection in August–September before seed dispersal. Public-health grade output. EU 2030 compliance ready.
Live
✏️
Draw Any Field
Click corners on the satellite map, double-click to finish. Name it, save it. Automatic hectare calculation from polygon geometry. No GIS knowledge needed.
Live
Technology Stack

Built on real science.
Validated like real science.

AgriGuard isn't a thin wrapper around free satellite imagery. Six layers of agricultural intelligence — engineered over four years, trained on EU reference data, transparently reported. Every pixel is auditable.

🌾
Vegetation Atlas
A living catalogue of crop spectral signatures across the EU growing season. Each entry contains the phenological trajectory, NDVI/NDRE curves, expected harvest window, and reference parcels. Continuously expanding via field-worker contributions.
Crop types10+ recognised
CoverageMulti-country EU
SourceLPIS + ground truth
🧠
BioGuardNet Neural Models
Multi-task neural architecture: classification, regression and spectral-index prediction in a single forward pass. Trained on radar + optical satellite fusion. Regional model variants for local crop diversity and growing conditions.
ArchitectureMulti-task NN
InputsRadar + optical
OutputsCrop · Index · Risk
🔬
Ground Truth Catalog
Every reference parcel, every field-worker observation, every confirmed crop feeds back into the training pipeline. Tens of thousands of validated records across multiple EU LPIS datasets. Continuous learning loop — accuracy compounds.
Validated parcels50,000+
PipelineField → GT → Re-train
Update cycleContinuous
🛰️
Multi-sensor Fusion
Optical satellite imagery (10–20 m, 9 spectral indices) combined with radar satellite (cloud-penetrating, all-weather). The radar layer ensures monitoring continues through cloud cover — critical for harvest-window decisions and growing-season continuity.
OpticalS2 · 10 m / 9 indices
RadarS1 · cloud-resilient
CadenceSub-weekly
🌱
Plant Health Surveillance
Per-field NDVI / NDWI / NDRE / EVI2 time series with anomaly detection. Early-warning signals for drought stress, disease onset, nutrient deficiency. Detects deviations 7–14 days before visible field symptoms — when intervention is still cheap.
Indices tracked9 per pixel
Anomaly window7–14 days lead
AlertsEmail · API · Mobile
📊
Transparent Validation
Every model ships with a validation report: F1 scores, confusion matrices, ROC curves, per-class precision/recall. Every prediction carries a confidence value. No black-box outputs — only auditable, reproducible inference. Ready for regulatory review.
ReportingF1 · ROC · Confusion
ConfidencePer-pixel value
AuditabilityFull pipeline trace
What the platform recognises — explicitly.
No abstract claims. The full inventory of crops, plant-health conditions, weeds and growth stages BioGuardNet identifies on every satellite pass. Per-pixel confidence on every prediction.
🌾
Crop Types
12+ recognised
🌽 Corn / Maize 🌾 Wheat 🌻 Sunflower 🌿 Rapeseed 🌱 Soybean 🍀 Alfalfa 🌾 Barley 🌾 Oats 🥔 Potato 🍇 Vineyard 🍎 Orchard 🟤 Sugar beet + regional
💧
Plant Health Conditions
6 stress signatures
💧 Water stress 🌱 Nitrogen deficiency 🔥 Heat stress 🍂 Senescence anomalies 🍃 Canopy density ⚠️ Lodging risk + disease pressure
🌿
Weeds & Invasives
Specialised models
🌿 Ragweed (Ambrosia) 🌾 Johnson grass 🌱 Foxtail 🟡 Wild mustard 🟢 Thistle species + regional weeds
📈
Growth Stages
Phenology tracking
🌱 Emergence 🌿 Vegetative 🌸 Flowering 🌾 Ripening ✂️ Harvest 🍂 Post-harvest + residue cover
The catalog grows with the data. Every new ground-truth submission expands what BioGuardNet recognises — new crops, new weed species, new stress patterns. The platform you license today gets smarter every quarter, automatically.
NDVI
Vegetation density & vigour — the canonical greenness indicator.
NDRE
Red-edge sensitivity — early stress & nitrogen status.
NDWI
Plant water content — drought stress detection.
EVI2
Enhanced vegetation — saturated-canopy correction.
LAI
Leaf Area Index — biomass & canopy structure.
SAVI
Soil-adjusted vegetation — sparse-canopy reliability.
CRE
Chlorophyll red-edge — photosynthetic activity.
PSRI
Plant senescence reflectance — maturity tracking.
BSI
Bare Soil Index — fallow detection & soil exposure.
Inputs
🛰️ Optical satellite
📡 Radar satellite
🗂️ LPIS ground truth
🌤️ Weather time-series
→
Processing
BioGuardNet
Multi-task neural network
S1 + S2 fusion · regional variants
→
Outputs
🌾 Crop classification
🌱 Health indices
⚠️ Risk surfaces
📊 Confidence scores
How AgriGuard Works

From raw imagery to decisions you can act on.

Nine stages — every parcel, every season, automatic. The same pipeline runs across Hungary, Romania, Italy, Poland — country-scale, ten-meter resolution, sub-weekly revisit.

Step 01 · AreaScan

Define your area.

Open the satellite map. Click corners to draw any region — a single field, a co-op, an entire county. Pick a pre-trained model. Set the year and the spatial resolution.

  • Polygon-based field definition · auto-hectare calculation
  • 200 m grid for region-scale · 10 m for precision parcels
  • No GIS knowledge required
01
AreaScan: drawing a polygon on the satellite map and selecting a model
Drawing · AreaScan
GT Sample Catalogue: 844K parcels across 22 plant species, satellite-validated training data, full crop rotation and feature extraction view
GT Catalogue · 844K parcels
02
Step 02 · GT Catalogue

Browse the truth catalogue.

844,546 reference parcels. 22 plant species. 4,810,000 hectares. Every parcel becomes a self-contained training card — coordinates, plant species, year, full spectral history. Click any parcel to see crop rotation, satellite scenes, and the 8 indices fed into BioGuardNet.

  • 844,546 reference parcels · 22 plant species · 2 countries
  • 9 spectral indices × 3 phenological phases = 54 features per card
  • LPIS-validated · year-specific · cloud-filtered (best record selection)
Step 03 · Pipeline

Extract features at scale.

The Feature Extraction Pipeline processes thousands of cards in queue. Each card gets satellite scene matching, cloud-masking, index computation, and multi-phase aggregation. Throughput is parallelised — country-scale runs complete in hours, not weeks.

  • Auto-orchestrated · resumable · per-card observability
  • Live status: Running · Queued · Done · Errors
  • 10 cards processed in parallel by default
03
Feature Extraction Pipeline: 11 cards being processed, RUNNING and QUEUED states visible
Pipeline · Live
BioGuardNet Training Studio: 6 ground-truth cards converging into a single neural model
BioGuardNet · Training Studio
04
Step 04 · Training Studio

Train a model on your region.

Select training cards from the catalogue. Connect them to BioGuardNet — a multi-task neural network combining a Random Forest backbone with an MLP classifier head. The studio renders the data flow live; when six or more cards are ML-Ready, training kicks off.

  • Architecture: RF(100) + MLP(64→32→C)
  • Multi-task: classification + regression + index prediction
  • Region-specific models for local crop diversity
Step 05 · Inference

Apply the model.

A four-stage pipeline runs in your browser: EU/LPIS truths are loaded, satellite samples are paired, BioGuardNet-H runs inference, and results validate against ground truth. Per-pixel confidence on every prediction.

  • EU/LPIS · Satellite × LPIS pairing · NN inference · validation
  • Live truth-set comparison · accuracy on every run
  • Output: crop class, plant health, confidence score
05
EU/LPIS Truths to BioGuardNet-H Inference: 4-stage pipeline with truth set validation
BioGuardNet-H Inference
Multi-index time series: NDVI, NDMI, PSRI, GNDVI across 2 years for 5 plant types
Time series · Live
06
Step 06 · Time series

Read the season — visually.

Every parcel becomes a multi-index time series. NDVI (vegetation), NDMI (moisture), PSRI (senescence), GNDVI (chlorophyll). Stress events, harvest dates and pest pressure surface as anomalies the eye reads instantly.

  • 97 satellite scans · 2-year window · 4 indices overlaid
  • Plant-type colour legend · per-parcel zoom
  • Outputs feed FieldGuard, Public Health surface, CAP report
Step 07 · Vegetation Atlas

Zoom into any parcel.

Click any field on the map. The Vegetation Atlas opens a parcel-scoped dashboard: NDVI health %, moisture index, ragweed risk score, recommended next steps. Every index plotted across the season — current scan, latest value, trend, status flag.

  • Live AI analysis · NDVI health · NDMI moisture · PSRI senescence
  • NDVI vegetation map with risk colour scale (red→amber→green)
  • Plant-type legend · 6 species · ambrosia surfaced separately
  • "Recommended Steps" — automated agronomic guidance
07
Vegetation Atlas: parcel-level dashboard with NDVI health, moisture, ragweed risk, index comparison and recommended steps
Vegetation Atlas · Per-parcel
Step 08 · Weather Intelligence

Read the climate, not just the calendar.

Every parcel gets its own weather profile — annual precipitation, max/min/mean temperature, sunshine hours, wind, evapotranspiration. Two-year baseline plus the current season. Correlations to NDVI computed automatically: precipitation, temperature, sunshine, ET₀ — each linked to vegetation response.

  • Monthly precipitation · temperature curves · wind · ET₀
  • Seasonal summaries: Spring · Summer · Autumn · Winter
  • Weather ↔ Vegetation correlations — 0.611 temp→NDVI, 0.822 sunshine→NDVI
08
Weather Dashboard: 2-year climate profile, monthly charts, seasonal summary cards, weather-vegetation correlations
Weather · 2-year baseline
10-day Forecast with Agrometeorological Alerts: irrigation considerations, soil moistening, 7-day water balance
10-day Forecast · Agro alerts
09
Step 09 · Forecast & Alerts

Decide tomorrow today.

A 10-day weather forecast for every parcel — temperature, precipitation, cloud cover, wind. Layered with agrometeorological alerts the agronomist can act on: high evaporation flags, irrigation considerations, soil moistening windows. A running 7-day water balance tells you whether your field is in deficit.

  • 10-day forecast · Open-Meteo · ERA5 reanalysis + GFS
  • Irrigation alerts · soil moistening windows · evaporation flags
  • 7-day water balance: precipitation − ET₀ = parcel status
Model Transparency · Explainable AI

Why does the AI think this is corn?

Every BioGuardNet decision is explainable. For every plant class, every season, the model surfaces which spectral bands and indices drove the call — and how strongly. No black box. No "the model said so."

BioGuardNet Feature Vector matrix: 17 spectral bands × 3 phenological phases, importance-coloured for Maize. Summer NIR, SWIR, NDVI all rated 5 (Final-tier importance).
Feature Vector · Maize · 2024
📊
17 bands × 3 phases = 51 cells
For every plant species, the model reports importance scores across raw bands (Blue, Green, Red, Red-Edge, NIR, SWIR) plus engineered indices (NDVI, NDWI, NDRE, etc.) — for spring, summer, and senescence.
🎯
Importance-tiered: 1–5 scale
Cells coloured Low (navy), Medium (steel blue), Important (amber), Final (red). Summer NIR + SWIR + NDVI score 5 for Maize — exactly as agronomic literature predicts.
🔬
Validation paths visible
Each model traces back to its training data — country LPIS files, year ranges, parcel counts, validation status. ARiMR PL ✓, APIA RO pending, MePAR HU pending. Full provenance.
Public Health Layer

Same satellite scan.
Different consumer.

Once a country is being scanned for crop intelligence, the same imagery yields a second, higher-stakes product almost for free: a national ragweed-risk surface for public health authorities.

Ragweed pollen is the dominant outdoor allergen across Central Europe. AgriGuard's risk surface lets ministries and municipalities target intervention before flowering — turning a reactive complaint-driven process into a planned public-health programme.

77M
Europeans affected
by 2050 (CABI)
14d
Lead time before
peak pollen release
∞
Country-scale,
same scan cycle
Public sector inquiries →
RAGWEED RISK · LIVE SURFACE
AgriGuard ragweed risk heatmap across Hungary
National-scale risk surface from a single satellite scan cycle.
The same data feeds AgriGuard's farm-level ragweed dashboard.
How It Works

From field to insight
in 4 steps

1
Draw Your Field
Click corners on the satellite map. Double-click to finish. Name it and save.
2
Run Analysis
Select date range. Click "Start". Our pipeline fetches the best cloud-free satellite scene.
3
Get Indices
9 spectral indices calculated automatically. Ragweed risk scored. Crop type estimated.
4
Take Action
Export reports, dispatch field workers, log activities, share with stakeholders.
Who Uses AgriGuard

Built for everyone
in the agriculture chain

From a 200-hectare family farm to a national paying agency — the same platform, scaled to your role.

🌾
Large Farm Operators (200+ ha)
Monitor every field from one screen. Get ragweed alerts before peak season. Prove compliance with photographic satellite evidence. Know what your neighbour is growing.
€4–5/ha/year
🏢
Agricultural Consultants & Cooperatives
Manage hundreds of client fields from one dashboard. Generate branded reports. White-label the platform under your company name. Full API access.
€12/ha/year · Enterprise
🏛️
Paying Agencies & Government
Automated CAP subsidy cross-check: declared crop vs. satellite-detected crop. Scale to millions of hectares. Catch fraud before payment. LPIS-compatible.
B2G · Custom pricing
🛡️
Agricultural Insurers
Validate crop damage claims with satellite evidence. Check if reported drought stress matches NDWI / NDVI data. Reduce claim processing time by 30–35%.
Insurance · API
⚕️
Public Health Authorities
National ragweed-risk surface for ministries and municipalities. Prioritise mowing operations before flowering. Measurable year-on-year reduction.
B2G · Public health
🌿
CSRD-Obligated Companies
From 2025, listed companies must report biodiversity impact (ESRS E4). AgriGuard provides Natura 2000 compatible monitoring reports for sustainability disclosure.
CSRD · ESRS E4
Coverage

Built for Central Europe.
Ready for the world.

Satellite imagery covers the entire globe. AgriGuard is optimised for EU regulatory frameworks and CEE agricultural conditions — but works on any field on Earth.

🇭🇺 Hungary — Live
🇷🇴 Romania — Live
🇸🇰 Slovakia — Live
🇵🇱 Poland — Available
🇮🇹 Italy — Available
🇩🇪 Germany — Available
🇦🇹 Austria — Available
🇺🇦 Ukraine — Agricultural aid
🇪🇺 EU CAP-ready · everywhere
🌍 Global — satellite worldwide
Wheat field after harvest with rolled hay bales and mountains in the distance
From planting to harvest
One scan cycle.
Every field, every season, measured.
Hungary · Romania · Italy · Poland · & beyond
Pricing

Simple, honest pricing.
No hidden fees. Ever.

Start free. Scale as you grow. Large operations get a tailored deal — because one size doesn't fit all.

Climate FieldView   €20+/ha
·
Sentera   €15–25/ha
·
CropX   €18+/ha
·
AgriGuard   from €4/ha
Free Trial
€0
⏱ 1 month · Max 50 ha
No credit card required
  • Up to 50 ha
  • Satellite analysis
  • 9 spectral indices
  • AI crop detection
  • Ragweed risk map
  • — Export / API
  • — Weather forecast
  • — ML model runs
Start Free →
Farmer
€5/ha/yr
Up to 500 ha · billed annually
100 ha€500/yr
250 ha€1,250/yr
500 ha€2,500/yr
  • Up to 500 ha
  • Unlimited S2 queries
  • Full time-series + weather
  • 9 spectral indices
  • AI crop + ragweed detection
  • Crop rotation + GT log
  • CSV + GeoJSON export
  • — API access
  • — White label
Get Started →
Professional
€4/ha/yr
500–2,000 ha · billed annually
500 ha€2,000/yr
1,000 ha€4,000/yr
2,000 ha€8,000/yr
  • 500–2,000 ha
  • Everything in Farmer
  • API access
  • BioGuardNet ML model
  • FieldGuard mobile app
  • PDF report export
  • Multi-user access
  • Priority email support
💡 For agronomists: manage multiple clients under one subscription.
Sign Contract →
PARTNER PROGRAM
Enterprise · Government
Custom deal
2,000+ ha · paying agencies · public health
  • Unlimited ha & users
  • White label option
  • Full API + webhooks
  • Custom ML model training
  • CAP / LPIS compliance suite
  • Public-health surface (ragweed)
  • CSRD / ESRS E4 reports
  • Dedicated SLA
  • Revenue share option
🤝 For consultants & agencies: you set the price to your clients. We set the wholesale rate together.
Request Partner Deal →
All plans include satellite data · No setup fee · Cancel anytime
Prices in EUR excl. VAT · Larger volumes always negotiated — contact us
Get Started Today

Your fields are being
monitored from space.
Are you watching?

Draw your first field in 60 seconds. Get satellite analysis in 2 minutes. Free for 1 month.