// the take
Official Python client for Sentinel Hub's satellite imagery services, wrapping Process, Catalog, Batch, BYOC, Statistical, and OGC APIs plus direct AWS access to Sentinel-2 SAFE archives. For developers building remote-sensing or geospatial pipelines who need Sentinel/Landsat/MODIS/DEM data without hand-rolling HTTP calls and auth against Sentinel Hub.
Auth, token refresh, and rate-limit backoff are handled for you (download/rate_limit.py, download/sentinelhub_client.py) instead of you reimplementing 429 handling. One package covers a lot of ground - Process, Catalog, Batch, BYOC, Statistical, OGC, plus AWS SAFE readers - so you're not stitching together five separate clients. It's a real production dependency, not a side project: it's the core data layer under eo-learn, has CI with coverage tracking, and ships type hints (py.typed). The area-splitting utilities solve the genuinely annoying problem of tiling a large AOI into requests that fit Sentinel Hub's per-request size limits.
Most of the useful functionality - Process API, Batch, BYOC - only works against a paid Sentinel Hub or CDSE account; despite the imagery topics, this isn't a free-download tool and the pricing model isn't explained anywhere near the README. There are parallel *_cdse.ipynb notebooks alongside the originals for nearly every example, which means the library is mid-migration from classic Sentinel Hub auth to Copernicus Data Space Ecosystem auth - expect to spend time figuring out which flow applies to your account. No built-in concurrency for bulk downloads beyond the rate limiter, so pulling many tiles means building your own parallelism on top. Documentation is spread across readthedocs prose and a dozen notebooks with no single quick-start that tells you what a typical request costs against your quota before you run it.
We read the README, the directory tree and a key source file in September 2026. Nobody pays to be listed. How we judge →