Roadmap · 2026 → 2027

Every feature, its own little world.

Below are seven scenes — each one a tabletop diorama, telling the story of a feature.

01 · SIMULATEQ2 · 2026

Watch the forest grow — before you plant a single seed.

Pick a parcel, draw a planting plan, and the simulation plays the next 20 years back to you in 40 seconds.

> STORY · Farmer draws 6 ha on the map. The table lights up. Saplings sprout in grid rows, the sun arcs overhead, canopy fills in by year 8, carbon credits accrue in real time.
  • Species-accurate growth curves (yield, canopy, carbon)
  • Monthly timeline slider · 0 → 20 years
  • Cost, survival rate & carbon credit forecasts
Engine · Opus 4.7 + SoilGrids3D · WebGL
LIVE GROWTH · 36 saplings
YEAR 08 / 20
AREA 6.2 ha · ZONE A
CO₂ +2,480 kg
ZONES · 4 ACTIVE
DRIP · 920 m
SAPLINGS · 1,240
SCALE 1:2000
02 · DESIGNQ2 · 2026

Your nursery, auto-designed.

Select a parcel. Get back a full nursery plan — irrigation lanes, species zones, access roads — tuned to soil, slope and shade.

> STORY · A 24 ha field becomes 4 zones in 8 seconds — almond here, pistachio there, poplar along the windbreak. Drip lines glow blue where water flows.
  • Zone-aware layout (dry, wet, slope, shade)
  • Drag-to-edit plots · species mix per zone
  • Export to CAD, KMZ, or printable A1
Region · UZBEKISTANScale · 1:2000
03 · MEASUREQ3 · 2026

The satellite scans. Every tree gets counted.

A satellite passes overhead. In the next frame, every tree in your parcel has a species, a health score, and a bounding box.

> STORY · Sentinel-2 fires at 10:42 UTC. By 10:43, 10,584 trees are identified, 6 species classified, 412 flagged as stressed.
  • 40+ species in the Central Asia catalog
  • Object detection on 10 cm drone mosaics
  • Confidence scoring · manual override UI
Model · CNN + SAM2F1 · 0.91
SCANNING · S-2
DETECTED 40 / 40
SPECIES · 3 CLASSES
F1 0.91
UAV · 120 m · 10 cm/px
S-2 · 10:42 UTC
CROP · WHEAT + ORCHARD
Δ WEEK +3.2%
04 · MEASUREQ3 · 2026

Two eyes in the sky. One unified census.

Sentinel-2 scans weekly at 10 m. Drones fly daily at 10 cm. We fuse both into a single, parcel-level picture.

> STORY · Drone mows the lawn pattern over 42 ha. Data streams rise from each strip. Upstairs, the satellite writes weekly deltas. Downstairs, parcels glow with this-week vs last-week change.
  • Automatic bounding-box tree detection
  • Crop-type classification (wheat, cotton, orchard…)
  • Change-over-time · weekly deltas
Tiles · 12,408Cost · $0.002 / ha
05 · REPORTQ4 · 2026

Every parcel tells its own story.

Each parcel gets a living document — what is planted, what changed, who is responsible, what the satellite saw last Tuesday.

> STORY · Reports float in stacks, re-writing themselves monthly. A ring fills as planting hits 84%. Another tells you canopy is at 62%, health at 91%. It's all ready for FAO, UNCCD or the EU.
  • Auto-refreshing PDF + shareable link
  • Carbon, water & yield dashboards
  • Regulator-ready formats (FAO, UNCCD, EU)
Output · PDF / PPTX / JSONAuto · Monthly
8 REPORTS · LIVE
AUTO · MONTHLY
PARCEL UZ-04-1223
PLANTED 84%
92 CONTRIBUTORS
CROPS · 128
LICENCE · CC-BY-SA
POOL · 6,420 rows
06 · COMMUNITYQ4 · 2026

Researchers feed the pool. The pool feeds the AI.

Field photos from agronomists in Tashkent, Fergana and Almaty flow into a shared research pool — which trains every detection model we run.

> STORY · 14 pins on a map. Each one is a researcher, a crop, a story. Lines rise to a central pool — the corpus. Every photo becomes training data for the next model.
  • OpenLayers-backed public map
  • Image + metadata ingestion (EXIF, GPS)
  • Licensing · CC-BY-SA · citation tracking
Base · OpenLayers 10Storage · S3 / pgvector
07 · INTELLIGENCEQ1 · 2027

Point your phone at a leaf. Get an answer in 2 seconds.

Bring your own Claude API key. Snap a leaf, a trunk, a root — we ask Opus 4.7 to identify the crop, spot diseases, and flag pests.

> STORY · Leaf floats in the hologram. Scan line sweeps top to bottom. Five amber spots light up — leaf-spot, 94% confidence. Octahedrons orbit the result: crop, disease, severity, action, cost.
  • 40+ diseases across 12 major crops
  • Pest ID with severity + recommended action
  • Works on phone, browser, or field tablet
Model · Opus 4.7 vision$10 ≈ 1,000 scans
SCANNING LEAF
OPUS 4.7
LEAF SPOT · 94%
$0.009 / scan
Bring your own key

$10 ≈ 1,000 crop & disease scans.

You bring your own Claude API key. We only package the request and render the result — transparent cost, full control stays with you.

Modelclaude-opus-4.7
Avg input2,400 tok
Avg output480 tok
Cost / scan≈ $0.009
$10 budget~1,000 scans