Three layers, one platform
ASTRA OS operates as a three-layer platform. Each layer builds on the one below, and each is independently valuable.
Unified Data Access
One API for all satellite data
Search, discover, and access satellite imagery from Sentinel-2, Landsat, and Planetary Computer through a single REST API. Consistent STAC metadata and COG output format regardless of the upstream provider.
- Unified catalog search across all providers (spatial, temporal, spectral)
- Automatic format normalization to Cloud-Optimized GeoTIFF (COG)
- STAC-compliant metadata for interoperability
- One API key, one set of docs, one integration
1"color: #6b7280"># Search across all providers at once2curl "https:"color: #6b7280">//astraos.cloud/api/v1/search?\3 bbox=-122.5,37.5,-122.0,38.0&\4 datetime=2025-01-01/2025-02-01&\5 cloud_cover_lt=20&\6 limit=10" \7 -H "Authorization: Bearer astra_your_key"Data Brokerage
Coming Soon — Commercial data marketplace
Purchase commercial imagery from Planet, Maxar, Airbus, Capella, and ICEYE through ASTRA OS. One contract replaces five vendor relationships. Intelligent source selection for the best available imagery.
- One contract replaces five vendor relationships
- Intelligent source selection (best imagery for time, location, budget)
- Volume discounts from negotiated provider agreements
- 15-25% margin on commercial data transactions
1"color: #6b7280"># Same API, commercial data included2scenes = astra.search(3 bbox=[-122.5, 37.5, -122.0, 38.0],4 datetime="2025-01/2025-02",5 collections=["planet-psscene", "maxar-wv"],6 resolution_lt=3 "color: #6b7280"># sub-3m resolution7)Analytics Primitives
Server-side processing via API
Pre-built processing functions that developers call via the API, eliminating the need to build common geospatial operations from scratch. NDVI, change detection, cloud masking, and more — all as callable endpoints.
- NDVI and vegetation index computation
- Change detection between time periods
- Cloud masking and atmospheric correction
- Multi-resolution fusion and co-registration
- Custom model hosting (deploy your own ML models)
1"color: #6b7280"># Compute NDVI — no local processing needed2job = astra.process(3 operation="ndvi",4 scene_id="sentinel-2-l2a:S2A_MSIL2A_20250115T...",5 bbox=[-122.5, 37.5, -122.0, 38.0]6)78"color: #6b7280"># Poll until complete9result = job.wait()10result.download("ndvi_output.tif")Why ASTRA OS
Provider-Agnostic
Add Sentinel today, Landsat tomorrow, Planet later. Zero API changes for users. The adapter pattern means new providers are invisible to your code.
COG + STAC Always
Every output is Cloud-Optimized GeoTIFF with STAC metadata. Even if the upstream provider delivers JP2 or SAFE format, we normalize it for you.
Built for Scale
Postgres-backed caching, async job processing, and controlled concurrency to upstream providers. The same architecture that scales from demo to production.