A product of

Aether AI

Every restaurant has cameras.Nobody's watching.

See how →

They record every table, every night.

Finished plates, empty glasses, raised hands. All of it, unseen.

Your cameras record every table, every night. Finished plates, empty glasses and raised hands all go unseen.

Introducing

DeepView F&B

Restaurant intelligence from the cameras you already have.

A product of Aether AI

Learn more →

The film

See it in 76 seconds.

From silent CCTV feeds to live nudges, your regulars and the bottom line.

00:00 / 01:16

DeepView F&B · by Aether AI

Better service starts with seeing every table.

Book a demo→

No table left waiting.

Computer vision on the cameras you already have, live nudges for your team, and reports for your managers.

Reads your whole floor

Tables seated, served, finished and cleared. Plates, glasses and raised hands, from the cameras you have.

See the live floor→

Nudges in real time

Staff get an alert the moment a plate is finished, a glass is empty or a hand goes up, so no one waits.

See what it catches→

Runs on-premise

Video is processed inside the restaurant and never leaves the building. No new hardware.

How it stays private→

Live analysis

Watch DeepView read the room.

One camera, the way DeepView sees it. Every person is found and tracked, every table is zoned, seats are counted and each change is logged the moment it happens, on the server inside the restaurant.

Sample footage

Seats taken over this clipClick to jump

Sample footage · people, tables and events are computed from this clip by detection and tracking

Live floor

Your whole floor, live.

Every table, the kitchen pass and the door, read from the cameras you already have. Tap a table to open its camera.

Harbor & Vine · Main Street Live

Live demo · sample dataFri 21:34

Tables in use17/24

Guests seated58across 4 sections

Wait for attention0:41avg, tonight

Turn time61 minavg, tonight

Free Seated Served Finished Check Needs clearing Resetting

Meet the cameras you already have. Now every one is on shift.

See the platform →

High camera view of a restaurant floor with tables marked seated, in mains, finished and free

CAM 01 · Main floor

Table turnover · seated to reset

Guests at a round table with a cup and a teapot marked as empty

CAM 07 · Table 12

Empty cups · drinks waiting

A guest raising her hand, tracked by pose points with a timer counting

CAM 02 · Table 9

Raised hand · time to attend

A finished plate on the kitchen pass with a waiting timer

CAM 05 · Kitchen pass

Plates waiting at the pass

Night terrace and sidewalk with people waiting at the door

CAM 04 · Door

Queue · walk-aways

A staff member wiping a table after guests left

CAM 06 · Dining

Tables left uncleared

A guest paying by card while the server waits

CAM 03 · Table 3

Check timing

A returning guest at his table, shown only under a code name

CAM 10 · Table 6

Returning guests · code name only

01 / 08

What DeepView watches for

Finished plates
Empty glasses
Raised hands
Table turns
Plates at the pass
Staff check-ins

Tables left uncleared
Walk-aways at the door
Section coverage
Check timing
Returning guests, by code name

Know your regulars. Not their names.

DeepView recognizes returning guests across your venues and shows them only under a code name, with their buying profile and what to offer. The data used for the match is deleted instantly and never stored.

Guest recognition · Harbor & Vine Main Street Returning guest

GUEST · JUNIPER-27 Matched on the DeepView server inside the restaurant

Visits: 9 across 3 Harbor & Vine venuesMain StreetTerracePier
Usually orders: Grilled seabass · Riesling
Spend profile: Premium

Suggest

Offer the chef's tasting menu

Name and personal details hidden
Match data deleted, never stored
Processed on-premise

The bottom line

Eight minutes a table adds up.

Every shift becomes a report. In this worked example, a faster turn at the dinner peak frees up tables you can seat again. Move the inputs to see how it scales.

Illustrative example

24 tables 3-hour dinner peak $96 average check

01Turn time at peak

~~68 min~~60 min

8 min faster

02Peak capacity

+8 tables a night

+0.353 turns per table × 24 = 8.47, counted as 8

03Over a year

× 360 nights

8 tables × $96 average check
= $768 a night

Added revenue · 8 tables × $96 × 360 nights

$276,480a year

Illustrative example, not customer results.

Every shift becomes a report

Harbor & Vine · Main Street Sample data

Covers per half hour

194covers so far · 212 forecast

Avg turn time

61min

Friday dinner, so far

Wait for attention

0:41avg

target under 1:00

Plates cleared within 2 min

92%

of finished plates

Check-ins per table

4.2avg

per seating

Runs on-premise

Your video never leaves the building.

Every camera streams to a server inside the restaurant, over the restaurant's own network. The video is read there, and it stays there.

No new hardware

DeepView F&B connects to the CCTV cameras the restaurant already has.

Processed in the restaurant

Video is read on a server inside the building and never leaves it.

Match data deleted instantly

When a returning guest is recognized, the data used for the match is deleted and never stored.

See every table.

Talk to Aether AI about putting DeepView F&B on the cameras your restaurants already have.

Book a demo → Watch the film →