Use Cases

Real solutions for real teams

See how teams across climate tech, agriculture, and insurance use ASTRA OS to eliminate data engineering and focus on what matters.

Climate Tech Startup

Climate tech startups building carbon monitoring, deforestation tracking, or wildfire prediction tools need imagery from multiple satellite providers. Each provider has different formats, APIs, authentication, and data delivery methods. A single integration can take weeks of engineering time.

Problem

Can't unify Sentinel + Landsat quickly. Engineers spend weeks writing custom ETL pipelines for each provider.

With ASTRA OS

One search → one API → done. Query both providers with a single request, get normalized COG output.

Days of integration become one afternoon.

climate_startup.py

1"color: #6b7280"># Before ASTRA OS: weeks of integration per provider
2from sentinelsat import SentinelAPI
3from landsatxplore import EarthExplorer
4"color: #6b7280"># ... custom format conversion, auth, normalization
5
6"color: #6b7280"># With ASTRA OS: one afternoon
7import astra
8
9scenes = astra.search(
10    bbox=[-62, -4, -60, -2],  "color: #6b7280"># Amazon region
11    datetime="2025-01/2025-02",
12    cloud_cover_lt=15
13)
14
15"color: #6b7280"># Both Sentinel-2 and Landsat results, normalized
16for scene in scenes:
17    print(f"{scene.provider}: {scene.datetime} ({scene.gsd}m)")
18    cog_url = scene.assets[&"color: #6b7280">#39;nir&#39;].url  # Always COG

Agricultural Analytics

Agricultural analytics companies need frequent optical imagery for crop monitoring and historical data for yield prediction. Sentinel-2 provides 5-day revisit for current conditions, while Landsat's 40-year archive enables long-term trend analysis. Building a unified time series across both requires significant data engineering.

Problem

Need frequent optical imagery + historical data for crop monitoring and yield prediction.

With ASTRA OS

Sentinel-2 for current 5-day revisit + Landsat archive for 40-year trend analysis. Unified time series through one API.

Seamless multi-source time series without data engineering.

agri_analytics.py

1import astra
2
3"color: #6b7280"># Build a unified time series across providers
4aoi = [10.5, 47.0, 11.5, 48.0]  "color: #6b7280"># Agricultural region
5
6"color: #6b7280"># Sentinel-2: recent high-frequency data
7recent = astra.search(
8    bbox=aoi,
9    datetime="2024-06/2024-09",
10    collections=["sentinel-2-l2a"],
11    cloud_cover_lt=20
12)
13
14"color: #6b7280"># Landsat: long-term historical trend
15historical = astra.search(
16    bbox=aoi,
17    datetime="2015-06/2024-09",
18    collections=["landsat-c2-l2"]
19)
20
21"color: #6b7280"># Same NDVI computation works on both
22for scene in recent + historical:
23    ndvi = scene.process("ndvi")
24    "color: #6b7280"># Unified time series, no format differences

Insurance / Risk Assessment

Insurance and reinsurance companies need satellite imagery for catastrophe modeling, property risk assessment, and claims verification. The challenge is discovering what imagery is available for a given location across all providers — and getting it in a consistent format for analysis.

Problem

Fragmented data access, manual discovery across providers. Slow underwriting analysis.

With ASTRA OS

Centralized catalog + consistent metadata. Search any AOI, get all available imagery ranked by relevance.

Faster underwriting with automated imagery discovery.

insurance_risk.py

1import astra
2
3"color: #6b7280"># Property risk assessment for underwriting
4property_aoi = [-90.1, 29.9, -89.9, 30.1]  "color: #6b7280"># New Orleans
5
6"color: #6b7280"># Search all available imagery for the area
7all_scenes = astra.search(
8    bbox=property_aoi,
9    datetime="2024-01/2025-01",
10    limit=50
11)
12
13print(f"Found {len(all_scenes)} scenes from "
14      f"{len(set(s.provider for s in all_scenes))} providers")
15
16"color: #6b7280"># Get pre/post event imagery for claims
17pre_event = astra.search(
18    bbox=property_aoi,
19    datetime="2024-08-01/2024-08-25"
20)
21post_event = astra.search(
22    bbox=property_aoi,
23    datetime="2024-09-01/2024-09-15"
24)
25
26"color: #6b7280"># Automated change detection
27delta = astra.process(
28    operation="change_detection",
29    before=pre_event[0].id,
30    after=post_event[0].id
31)

Ready to start building?

Get started with the free tier — 5,000 API calls per month, no credit card required.

1"color: #6b7280"># Before ASTRA OS: weeks of integration per provider 2from sentinelsat import SentinelAPI 3from landsatxplore import EarthExplorer 4"color: #6b7280"># ... custom format conversion, auth, normalization 5 6"color: #6b7280"># With ASTRA OS: one afternoon 7import astra 8 9scenes = astra.search( 10 bbox=[-62, -4, -60, -2], "color: #6b7280"># Amazon region 11 datetime="2025-01/2025-02", 12 cloud_cover_lt=15 13) 14 15"color: #6b7280"># Both Sentinel-2 and Landsat results, normalized 16for scene in scenes: 17 print(f"{scene.provider}: {scene.datetime} ({scene.gsd}m)") 18 cog_url = scene.assets[&"color: #6b7280">#39;nir'].url # Always COG

1import astra 2 3"color: #6b7280"># Build a unified time series across providers 4aoi = [10.5, 47.0, 11.5, 48.0] "color: #6b7280"># Agricultural region 5 6"color: #6b7280"># Sentinel-2: recent high-frequency data 7recent = astra.search( 8 bbox=aoi, 9 datetime="2024-06/2024-09", 10 collections=["sentinel-2-l2a"], 11 cloud_cover_lt=20 12) 13 14"color: #6b7280"># Landsat: long-term historical trend 15historical = astra.search( 16 bbox=aoi, 17 datetime="2015-06/2024-09", 18 collections=["landsat-c2-l2"] 19) 20 21"color: #6b7280"># Same NDVI computation works on both 22for scene in recent + historical: 23 ndvi = scene.process("ndvi") 24 "color: #6b7280"># Unified time series, no format differences

1import astra 2 3"color: #6b7280"># Property risk assessment for underwriting 4property_aoi = [-90.1, 29.9, -89.9, 30.1] "color: #6b7280"># New Orleans 5 6"color: #6b7280"># Search all available imagery for the area 7all_scenes = astra.search( 8 bbox=property_aoi, 9 datetime="2024-01/2025-01", 10 limit=50 11) 12 13print(f"Found {len(all_scenes)} scenes from " 14 f"{len(set(s.provider for s in all_scenes))} providers") 15 16"color: #6b7280"># Get pre/post event imagery for claims 17pre_event = astra.search( 18 bbox=property_aoi, 19 datetime="2024-08-01/2024-08-25" 20) 21post_event = astra.search( 22 bbox=property_aoi, 23 datetime="2024-09-01/2024-09-15" 24) 25 26"color: #6b7280"># Automated change detection 27delta = astra.process( 28 operation="change_detection", 29 before=pre_event[0].id, 30 after=post_event[0].id 31)