Product How It Works Status Privacy Founder FAQ Work with Us Book Pilot Assessment
A guest at Table 7 waiting, plates pushed aside
F1 · 19:35 SATURDAY

Right now, somewhere in your restaurant,
a guest is being forgotten.

The floor is alive. Your team is moving. And one table — you'll never know which one — is about to give up.

Multi-camera CCTV grid of a restaurant
F2 · SIX CAMERAS · ONE NIGHT

Six cameras. Twelve tables.
Three staff. Nobody sees everything.

Your CCTV records what happened. Your manager sees what's in front of them. The gap between the two is where customers are lost.

Empty Table 7 with abandoned plates and folded receipt
F3 · 19:53 — 18 MINUTES LATER

You don't lose customers in reviews.
You lose them in moments nobody saw.

They sat. They waited. They paid. They left. The floor never noticed. The review will arrive on Monday.

Sangati alert on Table 7 with phone notification
F4 · A DIFFERENT SATURDAY

This time, someone knows.
Within seconds.

Same camera. Same kind of moment. But the floor captain's phone just lit up — and Table 7 is no longer alone.

Phone showing Sangati dashboard, waiter saving Table 7
F5 · RESOLVED · 19:38

Even when you can't be there.
Sangati tells your team where to be.

No camera replacement. No POS swap. No proprietary device. Patterns flow up — never video.

Sangati.

The prediction layer for physical hospitality.

Book a pilot assessment
The Problem

Your POS explains what sold. It doesn't explain what went wrong on the floor.

Most dine-in restaurants operate with a visibility gap between the kitchen, the floor, and the people responsible for service recovery.

01

Missed service moments

A table waits too long for attention. A dirty table sits idle at peak. The manager finds out too late — or never.

02

Invisible floor bottlenecks

One zone overstaffed, another unmonitored. Hosts seat unevenly. No one catches it until service degrades.

03

No real-time operational visibility

POS is retrospective. Manager walks are intermittent. No continuous signal shows how floor ops are running right now.

What Sangati Monitors

What your team actually sees during service.

Live signals from your floor, routed to the right person at the right time.

Seating & occupancy

Zone-level table status, dwell time, and turn patterns across covered areas.

Wait time detection

Time from arrival to seating, and service attention gaps at occupied tables.

Floor coverage

Staff movement patterns and zone coverage to identify unattended areas.

Kitchen-to-table delays

Monitoring handoff bottlenecks between kitchen pass and floor delivery.

Drift from normal service patterns

Flags when service starts drifting from your restaurant's usual pattern — before anyone complains.

Role-based dashboards

Owner, Manager, Floor Captain, Kitchen, Host — each sees only what they need.

How It Works

Turn existing camera coverage into live operational visibility.

Four steps. No disruption to current operations.

STEP 01

Connect camera feeds

Taps into existing IP camera RTSP streams. Works with your current setup — no rewiring needed.

STEP 02

Process on-site

Inference runs locally on your existing infrastructure — a PC, server, or NVR-attached compute. No edge device to ship. Core detection stays on your premises.

STEP 03

Detect operational signals

Computer vision identifies occupancy changes, movement patterns, congestion, and service delays in covered zones.

STEP 04

Route to the right role

Alerts and dashboards push the right information to the right person — floor captain, kitchen, manager, or owner.

Product Preview

What the manager dashboard looks like.

Role-based views built for restaurant operators, not data analysts. This is a representative preview of the interface.

Sangati — Manager View
Floor Overview
Kitchen Status
Alerts
Staff Coverage
Daily Summary
Tables Occupied
18 / 24
75% occupancy · 3 tables pending clear
Avg Dwell Time
42 min
Baseline: 38 min · +10% above normal
Active Alerts
2
Table 7: 14 min no attention · Table 12: dirty 8 min
Attention Queue
3
Tables flagged for delayed service or clearing
PATIO
Normal
MAIN HALL
Busy
BAR AREA
Normal
PRIVATE
Alert

Representative interface preview for illustration. Actual dashboards vary by venue layout, configured zones, and role permissions.

Product Status

What's real today. What's next.

No vaporware, no roadmap passed off as reality.

Built & Internally Tested
  • Occupancy detection by zone
  • Dwell time tracking
  • Hybrid local-inference architecture
  • 5 role-based dashboard UX
  • Local-first event logging
  • CCTV RTSP feed integration
In Pilot Development
  • Service delay alerting
  • Kitchen pass monitoring
  • Staff zone coverage analysis
  • Role-based push notifications
  • Venue baseline learning
On Roadmap
  • Advanced venue baselines
  • Multi-zone alert refinement
  • Richer historical summaries
  • Multi-site dashboards
  • Role-specific reporting
Why This Matters

What changes when you have live floor visibility.

Faster response to service delays

Flags delays while the guest is still at the table — not after they've left a bad review.

Better zone visibility for managers

See which zones need attention without walking the entire floor continuously.

Tighter service consistency

Know when service is drifting from your restaurant's usual pattern — even when no one complains.

Less managerial guesswork

Replace instinct-driven floor walks with continuous live signals from every covered zone.

Sangati is built to reduce missed attention moments, improve table-flow visibility, and help operators intervene before service delays become lost covers, lower spend, or poor guest experience.

Why Not Existing Tools

Your POS, cameras, and floor walks each see a piece. None see the whole picture.

POS
CCTV
Manager Walks
Sangati
What it captures
What sold
Footage
What's visible now
Live operational deviations
When you see it
After close
On review
During the walk
Continuously during service
Can it catch a delay?
No
Only retroactively
If manager is there
While still recoverable

POS

CapturesWhat sold
TimingAfter close
Catches delays?No

CCTV

CapturesFootage
TimingOn review
Catches delays?Only retroactively

Manager Walks

CapturesWhat's visible now
TimingDuring the walk
Catches delays?If manager is there

Sangati

CapturesLive operational deviations
TimingContinuously during service
Catches delays?While still recoverable
Venue Fit

Built for a specific kind of restaurant.

Sangati isn't for everyone. Here's who gets the most value.

Best fit

  • Full-service dine-in restaurants
  • Multi-zone casual dining venues
  • Operators doing 50+ covers/day
  • Venues with existing IP camera infrastructure
  • Owners or managers wanting live floor visibility
  • Restaurants in Mumbai (for pilot phase)

Not ideal for

  • Delivery-only or cloud kitchens
  • Venues without usable camera coverage
  • Very small low-volume cafes
  • Operators expecting plug-and-play without calibration
  • Venues outside Mumbai during pilot
Pilot Process

What happens when you apply.

A structured process — not a sales pitch.

1

Qualification Call

We assess venue type, covers, and camera setup

2

Camera Assessment

Check coverage zones, angles, and RTSP compatibility

3

Edge Setup

Install and configure on-site processing device

4

Zone Calibration

Map floor zones to camera feeds and set baselines

5

Baseline Learning

System learns your venue's normal operating patterns

6

Operator Review

Review useful signals with you and iterate together

Privacy & Deployment

Your data stays in your restaurant.

Privacy is an architectural decision, not an afterthought.

On-premise processing

Inference runs locally on your existing infrastructure. No proprietary hardware. No video leaves your network.

No face recognition

Detects occupancy and motion patterns. Does not identify individuals.

No guest profiling

No personal data collection. No identity matching. No audio capture required.

Footage stays on-site

Raw footage does not need to leave your premises. Sangati extracts signals, not recordings.

Installation calibration

Zones and angles calibrated during setup. Works only within covered areas.

Standard camera support

Compatible with IP cameras outputting RTSP. No proprietary hardware lock-in.

Behind Sangati

Built by a restaurant operator,
not a software tourist.

Prasad Mehta

Founder & Managing Partner · Sangati Partnership

14 years in F&B operations. Built and ran two restaurant businesses, including UnCafe in Mumbai. Spent time working restaurant floors by choice. Building Sangati from floor pain, not from a pitch deck.

FAQ

Common questions.

Do I need new cameras?+
In most cases, no. Sangati works with existing IP cameras that output RTSP streams. Most modern CCTV systems support this. We assess camera compatibility during the qualification call.
Does Sangati replace my POS?+
No. Sangati layers on top of your existing systems. It adds operational visibility that sales data alone cannot provide.
What hardware is required on-site?+
Sangati installs as software on your existing infrastructure — a PC, server, or NVR-attached compute. No proprietary hardware, no edge appliance to ship. Inference runs locally on your hardware. Pricing during the qualification call.
Does it identify guests or staff by face?+
No. Sangati uses object detection to track occupancy and movement — not facial recognition. It cannot identify who someone is, only that a position is occupied or a zone has activity.
Does video footage leave my restaurant?+
Inference runs locally on your existing hardware. Sangati extracts operational signals — occupancy counts, dwell times, movement patterns — and sends only those patterns to our cloud. Raw footage stays on your network. We don't take custody of video, ever.
How long does setup take?+
Initial setup usually takes around 1-2 days depending on venue size, camera access, and calibration needs. Baseline learning improves over the first 1-2 weeks of operation as the system learns your venue's patterns.
What if camera angles are poor?+
Sangati works within covered zones. If a camera has poor angles or limited coverage, that zone will have reduced detection accuracy. We assess coverage during onboarding and are transparent about what each zone can and cannot detect reliably.
What's live today vs coming next?+
See the Product Status section for a transparent breakdown. Core detection and dashboard UX are built and internally tested. Service alerting and kitchen monitoring are in active pilot development. Advanced baselines, multi-site dashboards, and richer reporting are on the roadmap.
What does a successful pilot look like?+
A pilot is successful when the operator team finds the operational signals genuinely useful for daily service decisions — whether that's catching delays earlier, managing zones better, or having visibility they didn't have before. We define success together during the operator review phase.
Pilot Programme · Mumbai

PILOT INTAKE OPEN — MUMBAI VENUES

Book a founder-led pilot assessment

Limited pilot intake in Mumbai. We'll assess honestly whether Sangati fits your venue, camera setup, and operating model — no sales pitch.

Founder-led onboarding. Camera fit check included.

What you'll see in a walkthrough

Pilot intake · Mumbai

Reach out for a founder-led assessment.

Send venue details and we’ll respond within 48 hours with whether your setup is a fit and what a 30-day pilot would look like.

Send a pilot enquiry

We review every enquiry honestly. If your venue isn’t a fit, we’ll tell you.

Careers

Work with us.

Sangati is building the operational intelligence layer for India's restaurant industry. We're a small, fast-moving team based in Mumbai — looking for people who want to solve real problems for real operators.

Computer Vision Engineer

Object detection, tracking, and edge deployment. Experience with YOLO/ONNX and real-time video pipelines preferred.

MUMBAI · FULL-TIME

Full-Stack Developer

React, Node, Electron, Python. You'll build dashboards, APIs, and desktop tools for restaurant operators.

MUMBAI · FULL-TIME

Restaurant Ops Intern

Hospitality or F&B background. Help us design workflows that make sense on the floor, not just on a screen.

MUMBAI · INTERNSHIP

Don't see your role? We're always open to hearing from sharp people. Reach out at careers@sangati.tech