360° passive cue
A microphone ring listens in every direction at once — fully passive, no emissions, day or night, fog or glare. It detects and bears candidates the moment they make noise.
SENTIO — BY SIPSA LABS
Machines that hear, see, and act.
Sentio is an edge-AI sensing platform: a small, passive node that listens in 360°, looks where the sound is, confirms what it found, and hands a track to the systems you already run — on-device, no cloud in the loop.
FIRST MISSION
The drones that worry everyone are silent on radio.
A new class of small drone carries its control link over a fiber-optic thread instead of radio. It emits nothing — so the RF detectors and jammers most defenses rely on never see it. It is small, low, and fast enough to slip past cameras cued by nothing.
But propulsion is loud. A drone cannot mute its rotors. Passive acoustics catch the signature that signature-management hides everywhere else — and a camera, pointed by the sound, confirms it. That pairing is Sentio's first product: passive detection of RF-silent drones for the people protecting real sites.
HOW IT WORKS
Perceive. Confirm. Act.
One node, one loop, no operator babysitting. Each stage covers the failure mode of the one before it — sound cues in all directions at once, vision rejects what sound alone can't, and the result is an action, not a dashboard ping.
Look where it heard
A camera interrogates the cued bearing and confirms drone vs. not-drone — rejecting the mowers, generators, and gusts that fool any single sensor on its own.
Track out, hands off
Confirmed tracks publish in standard formats (TAK-compatible) to the C2, effector, or alerting stack you already operate. Sentio cues and confirms — it is a detection layer, not a weapon.
WHERE IT GOES
Defense first. Every unattended machine after — earned, not assumed.
The same node — ears, eyes, and an on-device brain — retrains for any environment where machines need senses and the cloud is too far away.
Defense & critical infrastructure
Counter-UAS detection for forces, bases, airports, stadiums, data centers, and the grid — the passive layer that catches what RF and imagery miss.
Agriculture
The watchful node for fields and herds: pests, intrusions, equipment health — sensed at the fence line, acted on locally.
Industry
Motors that report their own failing bearings. Lines that hear a jam before it cascades. Intelligence embedded in the machines themselves.
THE FIRST NODE
This geometry is real.
PRINTED IN-LAB · LIVE 3D OF THE ACTUAL ENCLOSURE · INTERNALS REPRESENTATIVE
WHERE WE ACTUALLY ARE
Built in the open. Claimed only when verified.
Sipsa Labs publishes what it can prove — a habit we built in our compression-research era and kept. Here is Sentio's real, current state:
- RUNTIMEWorking prototype. Full pipeline — acoustic cue → camera confirm → fused track → standard output — runs end-to-end today on commodity edge hardware, driven by simulated and recorded audio; the production mic ring is on order.
- VALIDATIONSimulation-validated. Detection, bearing, and multi-target separation hold in physics-based simulation. Field accuracy comes from field data — that campaign is the next phase, and we will publish what it shows.
- HARDWAREReference node in prototyping. Compact, passive, sub-watt-class sensing package; first enclosure printed and assembled in-lab.
- NOT YETNo fielded deployments, no performance claims. When we have field numbers, they will be published with their methodology — that is the standard here.
EARLY PARTNERS
We want to build this with the people who'll field it.
Integrators, site operators, and program offices: if RF-silent drones — or machines that need senses — are your problem, we want the conversation now, while the design is still being shaped by real constraints.
founder@sipsalabs.com