KML to Satellite Imagery,
Automated.

Upload a KML file, get analysis-ready satellite imagery back. TreeSight handles parsing, provider orchestration, clipping, and reprojection — so you don't have to.

Try Live Demo Request Early Access
Why TreeSight

The Problem With Manual Imagery Pipelines

📂

Fragmented Tooling

Parsing KML, searching providers, downloading assets, clipping, reprojecting — each step requires different tools and APIs.

⏱️

Slow Turnaround

Manual pipelines take hours to days. TreeSight's durable orchestrator typically completes in minutes for standard AOIs, with automatic retry and parallelism. Processing time varies with AOI size and cloud cover.

🔐

Multi-Tenant Blind Spots

Container-per-tenant isolation, tenant-scoped routing, and valet tokens keep data separated without complex infrastructure.

📐

Geometry Gotchas

Unclosed rings, coordinate normalisation, geodesic area calculation, buffered bounding boxes — TreeSight handles all of it correctly.

Capabilities

What You Get

KML polygon extraction (Fiona + lxml fallback)
Geodesic area and buffered bounding boxes
Pluggable provider registry (Planetary Computer, extensible)
Durable Functions orchestration with fan-out parallelism
Automatic retry with exponential backoff
Clip and reproject to target CRS
Try It

Live Demo

🗺️
New to KML files?

Learn how to draw an area of interest in Google Earth Pro and export it as KML — no GIS experience needed.

Read the Guide
API contract version mismatch. Demo submissions are disabled until the backend is updated.
📁
Drop a .kml or .kmz file here, or click to browse
Supports KML and KMZ files up to 5 MB
— or paste KML text —

Your data is automatically deleted after 30 days. Privacy Policy

○ Uploading KML…
○ Parsing & preparing AOIs
○ Searching imagery (NAIP + Sentinel-2)
○ Downloading, clipping & reprojecting
○ Weather, NDVI & mosaic caching
○ Complete

Pipeline Results

Before
After
Loading imagery…
Left: Right:
Registering mosaics with Planetary Computer…
Caching tiles…

🔍 Analysis

Scope: All areas (collective)

🌡 Climate Context

Temperature (°C)Precipitation (mm)

🌿 Vegetation Health Trend

Mean NDVIΔ drop ≥ 0.1

📊 Period Summary

How It Works

Pipeline Workflow

1

Upload KML

A KML file is uploaded to blob storage, triggering an Event Grid event.

2

Parse & Validate

Fiona (or lxml fallback) extracts polygon features, normalises coordinates, and auto-closes rings.

3

Prepare AOIs

Each feature becomes an Area of Interest with bounding box, geodesic area, centroid, and buffered search extent.

4

Search & Queue

Composite search: NAIP for high-res detail (US only) + Sentinel-2 for global temporal monitoring. Matching scenes selected and queued in parallel across all AOIs.

5

Download

Queued imagery assets are downloaded in configurable batches with retry and exponential backoff.

6

Clip & Reproject

Downloaded GeoTIFFs are clipped to AOI boundaries and reprojected to the target CRS.

Pricing

Simple, Transparent Pricing

Start free, upgrade when you need more. All plans include satellite imagery from Sentinel-2 and NAIP, weather correlation, and NDVI change detection.

Free
$0 /month
For individual researchers and evaluation
  • 5 AOI analyses total (beta)
  • 1 concurrent pipeline
  • Sentinel-2 (global) + NAIP (US) imagery
  • NDVI + weather overlay
  • Before/after comparison
  • 30-day data retention
  • Community support
Start Free
Pro
$49 /month
For small NGOs and independent advisors
  • 50 AOI analyses per month
  • 3 concurrent pipelines
  • Everything in Free, plus:
  • AI-generated insights
  • GeoJSON & CSV export
  • 90-day data retention
  • Email support
Get Pro
Team
$149 /month
For agritech companies and compliance teams
  • 200 AOI analyses per month
  • 10 concurrent pipelines
  • Everything in Pro, plus:
  • API access
  • Up to 5 team members
  • 1-year data retention
  • Priority support
Get Team
Enterprise
Custom
For large organisations and government agencies
  • High-volume AOI analyses
  • Dedicated infrastructure
  • Everything in Team, plus:
  • SSO / SAML integration
  • Custom integrations
  • SLA & audit logs
  • Dedicated account manager
Talk to Sales

All plans include access to Microsoft Planetary Computer imagery at no extra cost. Enterprise pricing scales with compute usage. Need more AOIs? Contact us to discuss high-volume pricing.

Use Cases

Built for Real-World Geospatial Workflows

A conservation team monitoring 140 forest concessions across Central Africa replaces a 3-day manual QGIS workflow with a 10-minute automated pipeline — and uses AI summaries to save hours writing donor reports.
Conservation
An agricultural advisory uses integrated NDVI + weather correlation for crop insurance assessments. One upload replaces three separate tools, with everything in a single dashboard.
Agriculture
An ESG compliance team uses before/after satellite evidence with quantified change metrics for EUDR reporting. The comparison view and hectare-level change detection provide audit-ready output.
ESG Compliance
Transparency

Data Sources & Methodology

TreeSight uses publicly available satellite imagery and environmental data. Here's exactly where the data comes from and how we process it.

Sources

Layer Source Resolution Coverage Frequency
Optical imagerySentinel-2 L2A via Planetary Computer10 mGlobal~5 days
High-res imageryNAIP via Planetary Computer0.6 mUS (CONUS) only~2–3 year cycle
WeatherOpen-Meteo Historical Archive~25 km gridGlobalDaily
Flood eventsUK Environment Agency / USGS streamflowPoint/areaUK, USBest-effort / varies
Fire hotspotsNASA FIRMS VIIRS/SNPP375 mGlobalBest-effort / varies

Note: Flood and fire layers are experimental and depend on third-party APIs and required access keys. Availability and recency may vary, and these layers can sometimes return no data.

How We Process It

Known Limitations

Questions

FAQ

What file formats are supported?
KML files with Polygon and MultiPolygon geometries. The parser handles both Fiona/GDAL and a pure-lxml fallback for environments without GDAL.
What satellite imagery providers are supported?
Microsoft Planetary Computer is the default provider, with composite search: NAIP for high-resolution detail (0.6 m, US coverage) and Sentinel-2 L2A for global temporal monitoring (10 m). The provider registry is pluggable — additional providers can be added by implementing the ImageryProvider interface.
How does multi-tenancy work?
Each tenant gets isolated input/output containers (e.g. acme-input, acme-output). The tenant ID is derived from the container name, and all pipeline outputs are scoped accordingly.
What happens if an imagery download fails?
Downloads are retried with configurable intervals and exponential backoff. Failed or timed-out downloads are flagged as partial in the pipeline summary, while successful downloads continue through fulfilment.
How large can KML files be?
The maximum KML file size is 10 MiB. For very large feature sets, the payload offloader automatically moves data to blob storage to stay within the Durable Functions 48 KiB history limit.
Is the demo data real satellite imagery?
The demo map shows real imagery from Microsoft Planetary Computer: a NAIP detail layer (~0.6 m) for high-resolution context, plus a Sentinel-2 L2A temporal series (10 m, bi-monthly) to track change over time. The pipeline uses the same composite search strategy in production.
What CRS are outputs delivered in?
By default, outputs are reprojected to EPSG:4326. You can specify a custom target CRS in the pipeline configuration.
How is the pipeline deployed?
Azure Functions on Container Apps with Durable Functions orchestration, Azure Blob Storage, Event Grid triggers, and a Static Web App for the frontend. All managed via OpenTofu.
What happens to my data?
Uploaded KML/KMZ files and analysis results are automatically deleted after 30 days. Server logs are retained for 90 days for debugging. We do not sell or share your data. See our Privacy Policy for full details.
Get Started

Request Early Access

Tell us about your organisation and use case. We'll be in touch.