INTRODUCING SKY GARUDA PROTOTYPE / 01

EYES IN
THE SKY.
AI MONITORING.
HUMANS
IN CONTROL.

Less noise. More context. Turn aerial awareness into coordinated public-safety response—with a human behind every decision.

▶ Explore the scenarios
✳Built for the moment when
seeing clearly changes what happens next.
AERIAL PERSPECTIVEPROMOTIONAL VISUAL
17.4155° N
78.4358° E
PROMOTIONAL VIDEO · NOT A LIVE FEED01 — 04 ↗
A BETTER
POINT OF VIEW ✳
Observe & FlagAI Alerts

Surface signals. Bring the evidence.

Illustrative workflow
Verify & RespondHuman Decision

Review the context. Own the call.

Operator approval first
Launch VideoMuted · Looping
AERIAL AWARENESS✳GEO CONTEXT✳HUMAN-LED RESPONSE✳EVERY DECISION TRACEABLE✳

01 / THE PLATFORM

One picture.
Fewer blind spots.

A drone feed alone is only part of the story. Bring aircraft, observed conditions and incident decisions into one shared workspace.

01
⌁

Know your fleet.

Track aircraft position, battery and observation freshness. Know when a signal is current—and when it isn’t.

FLEET + TELEMETRY
02
◉

Analyze the feed.

Locate anonymous face regions and flag weak video quality. People stay unnamed; operators review the diagnostics.

ANONYMOUS FACE ANALYSIS
03
⌖

Put it in context.

Review monitored zones and map-linked events. Keep location provenance and uncertainty in the picture.

ZONES + GEOSPATIAL
04
↗

Make the next call.

Inspect the evidence, acknowledge an incident and record an action. An alert starts a review, not an automatic response.

REVIEW + AUDIT
↳ PROTOTYPE, WITH RECEIPTS.

The current workspace uses simulated crowd observations and fixture zone locations. A separate internal worker supports video-quality checks and baseline person detection. Aerial accuracy and calibrated ground projection still need evaluation.

02 / BUILT FOR WHAT HAPPENS NEXT

Different situations.
Same clear head.

Good conditions. Bad signal. A crowded exit.
Explore the decisions behind the dashboard.

NORMAL FLOWILLUSTRATIVE PREVIEW
CONTROLLED VENUE
34synthetic count
✳
● All eyes on the contextNo threshold exceedance
N
↑

OBSERVE / UNDERSTAND

Stay aware. Stay measured.

Follow aircraft and zone observations in one view. Normal activity stays visible without turning every movement into an incident.

OPERATOR ACTIONMonitor the shared picture.
PLUS THE OPERATIONAL DETAILS THAT MATTER

Weather changesInspect official context and data age. Weather does not equal flight clearance.

Evidence reviewAcknowledge, dismiss or resolve with a decision note and an audit record.

Volunteer checkpointSeparate consent and manual review. No public face search or biometric matching.

Lost connectivityKeep stale inputs visible. Field offline buffering remains planned work.

03 / FROM SIGNAL TO DECISION

See it.
Check it.
Own the call.

Human judgment isn’t a fallback.
It’s the centre of the workflow.

  1. 01

    Observe the condition

    Bring aircraft telemetry and monitored-zone observations into the workspace. Surface freshness and source quality.

  2. 02

    Review the evidence

    Open the incident. Inspect the trigger, threshold, persistence and location provenance before deciding.

  3. 03

    Coordinate a response

    Acknowledge, dismiss or resolve. Record the action and reason. Exercise actions do not dispatch real teams.

  4. 04

    Leave a clear trail

    Review the local audit chain and export the exercise. Accountability stays part of the operation.

04 / NO BLACK BOX ENERGY

Powerful perspective.
Clear boundaries.

Trust starts with saying exactly what a system can do—and what it can’t.

↗

People make decisions.

No automatic enforcement. No inference of criminal intent. Every incident calls for operator review.

⌖

Location needs proof.

Aircraft GPS isn’t a person’s location. Imported airspace and fixture zones are labeled. Ground projection needs validated calibration.

◉

Sources stay honest.

IMD, SACHET and Bhuvan adapters need configuration and validation. Other agency services remain disconnected.

↳

Consent stays separate.

The checkpoint records consent and manual review. It stores no face images, templates or biometric scores.

CONTEXT ADAPTERSIMDNDMA SACHETISRO BHUVANConfiguration required · No partnership or endorsement implied

THE PRACTICAL STUFF

Before you take off.

Is this connected to a real drone?+

The included exercise uses synthetic telemetry. Signed HTTP gateway ingestion is implemented, but your authorized fleet still needs a protocol adapter. There is no aircraft-control feature.

Does the prototype analyze drone video?+

The Feed & face analysis view accepts private video uploads, configured storage API imports, registered recordings and authorized RTSP feeds. It runs anonymous face detection, image-quality checks and optional baseline person detection. The exercise counts remain synthetic. Aerial validation, continuous streaming, tracking and calibrated ground projection remain deployment work.

Can it identify people in public footage?+

No. Public face search and watchlist matching are not implemented. A separate volunteer checkpoint supports consent records and manual review only.

What can I try in the dashboard today?+

Run four exercise conditions, inspect aircraft freshness, edit zone thresholds, review incidents, inspect source adapters, manage volunteer consent, verify the local audit chain and export the exercise.

Is Sky Garuda production-ready?+

This is a local hackathon prototype. Agency authorization, verified data, real video evaluation, TLS, production identity and operational acceptance are separate deployment requirements.

THE NEXT MOVE IS YOURS.

Better awareness.
More considered action.

✳

Explore the local prototype. Keep the decisions human.