Sarvajña — Unified Geospatial Inferencing Platform

स

Sarvajña

Ask the Earth.
Get a decision.

STORM ONLINE · 8 SATELLITES · ISL ACTIVE · VIRTUAL CONSTELLATION

02 · Three Pathways

One platform, three ways in.

Whether you need raw analysis-ready data, a remote-sensing index, or want to deploy your own model on orbit, Sarvajña handles all three through a single unified back-end.

Pathway 01

Analysis Ready Data

Get pre-processed, calibrated, atmospherically-corrected imagery over your area of interest. L2A reflectance, COG-packaged, STAC-indexed.

$ sarvajna ard --aoi krishna_delta.geojson \
--sensor optical --resolution 10m --period 14d

→

Pathway 02

Indices Inference

Pick from NDVI, SAVI, NDWI, NBR and 40+ indices spanning vegetation, water, soil, burn, snow, urban, geology, SAR and hyperspectral domains.

$ sarvajna index --type NDVI --aoi vidarbha.kml \
--cadence weekly --output insights+chip

→

Pathway 03

GeoAI Conversion

Upload your geospatial model. Sarvajña optimises it, applies lossless compression, validates against STORM hardware, and deploys it in orbit.

$ sarvajna deploy ./ship_detector.onnx \
--target storm --preprocessing L1C --quantize int8

→

03 · Live Console

Configure once. Sarvajña handles the rest.

Six parameters. Six clicks. The same form drives all three pathways — and the back-end decides whether to run it on the ground, on STORM in orbit, or hybrid.

Active pathway: Analysis Ready Data

PATHWAY · 01

i Area of Interest

Krishna Delta
16.05°N – 16.55°N, 80.85°E – 81.40°E
◆ 4,820 km² · GeoJSON loaded

ii Sensor Type

Optical

SAR

Hyperspectral

Thermal IR

LiDAR

iii Resolution

VHR < 1m

HR · 1–5m

MR · 5–30m

LR > 30m

iv Revisit

Sub-daily

Daily

Weekly

Bi-weekly

v Orbit Type

LEO Sun-Sync

LEO Polar

MEO

GEO

Equatorial

vi Output Type

Insights only

Insights + chip

Insights + raw

Time-series stream

est. time ~ 38s est. cost ₹ 2,140 execution SURGE → STORM (edge)

◆

Execution trace

0.00s

Decision Delivered

query_id: qry_8a31f4_001

Insights — structured output

JSON

Image chips · top detections

3 CHIPS

04 · Virtual Constellation

One detection. Two satellites. No ground delay.

When SAT-01 flags an event, the onboard orchestration engine selects the best-positioned satellite based on orbital geometry and passes the cue directly via ISL. No ground command cycle is needed. Intelligence is delivered in minutes, not hours.

Auto-advances 2.5 s per step · ISL tip-and-cue workflow · 4-satellite virtual constellation

05 · Platform Preview

What mission operators see.

The Sarvajña mission control interface gives operators real-time awareness across the virtual constellation — active tasking, live detections, ISL health, and prioritised alert routing, all in one view.

sarvajna.skyserve.ai / mission-control

स

Sarvajña MISSION CONTROL

3 SAT ONLINE

4 ALERTS

ISL NOMINAL

SAT-04 PENDING

Alert Feed

CRITICAL · 14:32 IST

IUU Vessel Detected

18.2°N 73.5°E · Arabian Sea

Conf. 94% · SAT-01 · via ISL

WARNING · 14:18 IST

Oil Sheen Detected

19.1°N 72.8°E · Off Mumbai

Conf. 81% · SAT-02

INFO · 14:05 IST

Cyclone Low Forming

12.4°N 87.2°E · Bay of Bengal

Conf. 73% · SAT-01

INFO · 13:58 IST

Port Congestion Flag

JNPT · Navi Mumbai

Conf. 88% · SAT-03

ISL STATUS SAT-01 ══ SAT-02 ══ SAT-03 ── SAT-04 Last handoff: SAT-01 → SAT-02 · 4 min ago · Next: SAT-02 → SAT-03 in 8 min

06 · System View

Under the hood — what the user never has to touch.

Sarvajña is a seven-layer stack. Below, the architecture and the end-to-end data processing workflow that the platform runs on every request.

Seven-Layer System Architecture

v1.0 · MAY 2026

Ask the Earth.Get a decision.