CLI
The InSARHub CLI (insarhub) drives the whole pipeline — search, process, analyze — from a single command. This page shows the end-to-end workflow: which commands to run, and in what order. For every subcommand and flag, see the CLI Reference, or run --help on any command.
Workflow
graph
A[Search Scenes] --> B[Select Pairs];
B --> C[Interferogram Processing];
C --> D[Time-series Analysis];
D --> E[Post-Processing];
click A "../../advanced/cli_reference/#downloader" "insarhub downloader"
click B "../../advanced/cli_reference/#pair-selection" "insarhub downloader --select-pairs"
click C "../../advanced/cli_reference/#processor" "insarhub processor submit"
click D "../../advanced/cli_reference/#analyzer" "insarhub analyzer run"
click E "../../advanced/cli_reference/#utils" "insarhub utils"Every run is the same four stages — search & select → process → analyze → post-process — and differs only in which processor/analyzer backend does the work. Each command writes an insarhub_config.json to the workdir and reloads it on later runs, so after the first call you only pass what changed.
HyP3
Cloud processing — no local tools needed; interferograms are generated by ASF HyP3.
# 1. Search and select pairs
insarhub downloader -N S1_SLC \
--AOI -113.05 37.74 -112.68 38.00 \
--start 2020-01-01 --end 2020-12-31 --stacks 100:466 \
-w /data/bryce --select-pairs
# 2. Submit interferograms to HyP3
insarhub processor -N Hyp3_S1 -w /data/bryce submit
# 3. Check job status (re-run until all SUCCEEDED)
insarhub processor -N Hyp3_S1 -w /data/bryce refresh
# 4. Download completed outputs
insarhub processor -N Hyp3_S1 -w /data/bryce download
# 5. Time-series analysis
insarhub analyzer -N Hyp3_Mintpy_SBAS -w /data/bryce run
ISCE2_S1
Requires a local ISCE2 install (or --container). SLC .SAFE files are downloaded first.
# Search and download SLC scenes + orbits
insarhub downloader -N S1_SLC \
--AOI -113.05 37.74 -112.68 38.00 \
--start 2020-01-01 --end 2020-12-31 --stacks 100:466 \
-w /data/p100_f466 --select-pairs --download -O
Process the stack, then run the time-series (add --dry-run to the submit to check paths first):
# Export velocity to GeoTIFF
insarhub utils h5-to-raster -i /data/p100_f466/mintpy/geo/geo_velocity.h5
GMTSAR_S1
Requires a local GMTSAR install (or --container). SLC .SAFE files are downloaded first; the DEM is auto-downloaded from the footprint.
# Search and download SLC scenes + orbits
insarhub downloader -N S1_SLC \
--AOI -113.05 37.74 -112.68 38.00 \
--start 2020-01-01 --end 2020-12-31 --stacks 100:466 \
-w /data/gmtsar --select-pairs --download -O
Generate interferograms, then run the MintPy time-series:
# Export velocity to GeoTIFF (GMTSAR output is already geocoded)
insarhub utils h5-to-raster -i /data/gmtsar/gmtsar_mintpy/velocity.h5
ISCE3_Burst + Dolphin
Requires the ISCE3 / COMPASS / dolphin stack (or --container). Produces geocoded velocity directly — no h5-to-raster needed.
# Search and download Sentinel-1 bursts (assembled into .SAFE)
insarhub downloader -N S1_Burst \
--AOI -106.06 40.34 -105.70 40.58 \
--start 2025-01-08 --end 2025-04-26 \
-w /data/p56 --select-pairs --download
Run the burst pipeline (dem, tec, cslc, static, ifg, stitch, unwrap, los), then the Dolphin time-series:
The velocity lands at /data/p56/timeseries/velocity.tif (already GeoTIFF — no h5-to-raster needed).
NISAR_GSLC + Dolphin
Requires the ISCE3 / dolphin stack (or --container) — same image as ISCE3_Burst. NISAR GSLC is already geocoded, so there is no coregistration/geocoding: the stack goes straight into dolphin. ISCE3_NISAR crops each GSLC to the AOI first, so a small AOI stays fast and light even though the source frame is huge.
# Search and download NISAR L2 GSLC frames (one geocoded frame per date)
insarhub downloader -N NISAR_GSLC \
--AOI -113.08 37.68 -112.58 38.07 \
--start 2025-11-01 --end 2026-09-01 \
-w /data/p77 --download
Run the dolphin pipeline (ifg, stitch, unwrap), then the Dolphin time-series:
The velocity lands at /data/p77/timeseries/velocity.tif. To process the whole geocoded frame instead of the AOI window, add --process_full_extent (needs a large-memory host).
Full flags
For every command's options, run insarhub <command> --help or see the CLI Reference. To run any backend without a local install, add --container <image> — see Container Execution.