Welcome

We blend strategy, design and engineering to build intelligent systems that help businesses in Saudi Arabia and the GCC grow and stand out.

Simat

Multi-branch restaurant platform: POS, kitchen, delivery, stock and demand forecasting

  • A Shfrah product
  • Food & retail
  • AI
  • SaaS
  • At a glance

Sector

Food & retail

Platforms

Web · iPad · Android · iOS

Modules

9

Integrations

7

Screens

4

Total duration

24 weeks

Simat brings every branch and the central kitchen onto one system: tablet POS that keeps selling when the internet drops, a kitchen display, delivery-app orders in one inbox, and stock tied to recipes and food cost. It forecasts demand per branch and hour to propose a prep and purchasing plan every morning, and shows owners the profit of each branch day by day.

Simat: screen 1
Owner dashboard: sales by channel, food cost, top items and branch ranking.
  • Challenge

  1. 01

    Delivery-app orders arrive on a separate tablet per app, cashiers re-key them by hand, and some get lost at peak time.

  2. 02

    Real food cost is unknown: recipes are not tied to stock and waste goes unrecorded, so the gap only shows at the month-end count.

  3. 03

    Prep and purchasing are left to the supervisor to estimate, so waste climbs on quiet days while items sell out at peaks, during Ramadan and in high season.

  4. 04

    Changing a price or pausing an item means logging into every delivery app one by one, and owners see branch profitability too late.

Who it is built for

  • Owner

    Daily profit per branch and channel, and food cost, on the phone

  • Branch manager

    A ready prep and purchasing plan each morning, plus waste and stock counts

  • Cashier

    Fast checkout that keeps working offline, with every order on one screen

  • Head chef

    Orders by station, with prep time for every dish

  • Central kitchen & purchasing

    Branch transfers and supplier orders driven by the forecast

  • Modules

What the system does

Connected modules working on one set of data.

  • 01

    Point of sale

    Tablet POS on iPad and Android that works offline and syncs on reconnect, with mada and Apple Pay

  • 02

    Kitchen display

    Routes items to stations, times every order and alerts past the target time

  • 03

    Delivery inbox

    HungerStation, Jahez, Keeta, ToYou and other orders on one screen, straight to the kitchen

  • 04

    Menu sync

    One menu with prices and availability per channel and branch, published to every app in one click

  • 05

    Inventory & recipes

    Deducts ingredients by recipe on every sale, logs waste with a reason, and compares actual with theoretical food cost

  • 06

    Central kitchen

    Batch production and transfers to branches with lot numbers and expiry dates

  • 07

    Forecast & prep plan

    Item sales forecast per branch and hour, and a daily prep and purchase plan to approve

  • 08

    Loyalty & offers

    Points, vouchers and time- or branch-based offers, recognising guests by mobile number

  • 09

    Invoicing & P&L

    Simplified invoices compliant with ZATCA Fatoora, and a P&L for every branch

Where the AI works

  • AI · 01

    A gradient-boosted time-series model (LightGBM) forecasts each item per branch and hour from sales history, weather, holidays, paydays, Ramadan, seasons and match nights.

  • AI · 02

    A planner turns the forecast, through recipes, into prep quantities per station and purchase orders that respect stock on hand, supplier lead time and shelf life.

  • AI · 03

    Anomaly detection on waste and on actual versus theoretical usage points the owner to a specific branch or shift.

  • AI · 04

    Prep-time estimation per order from its items and station load, sent to the delivery app so the courier arrives on time.

  • Screens

The system as the client's team and customers use it.

Simat: screen 2
Kitchen display: in-store and app orders split by station, with a timer on every order and alerts past the target time.
Simat: screen 3
Prep plan: the two-hourly demand forecast and the day’s drivers, then proposed prep quantities and purchase orders to approve.
Simat: screen 4
Owner app: yesterday across all branches, with a stock-variance alert before it piles up.
  • Architecture

Independent services, in-Kingdom data, and integrations over documented APIs.

Channels

  • POS tablet: iPad & Android · offline-first
  • KDS screens: Station displays
  • Owner app: iOS & Android
  • Back office: Web admin

Edge

  • API gateway: Auth & rate limits
  • Sync service: Offline sync & conflict resolution
  • Realtime hub: WebSocket to the kitchen

Services

  • Order service: Orders from every channel
  • Menu service: Prices & availability per channel
  • Delivery aggregator: Delivery apps
  • Inventory service: Recipes, stock & waste
  • Finance & invoicing: Invoices & P&L

AI

  • Demand forecast: Per branch, item & hour
  • Prep planner: Prep & purchase plan
  • Variance detection: Waste & stock variance

Data

  • PostgreSQL: Orders, stock & menus
  • Kafka: Sales & kitchen events
  • ClickHouse: Analytics & training data

Integrations

  • HungerStation · Jahez: Partner APIs
  • Keeta · ToYou: Partner APIs
  • ZATCA Fatoora: Simplified invoices
  • Payment provider: mada & Apple Pay
  • WhatsApp Business: Receipts & offers

Connections

  • POS tablet to Sync service: sync
  • KDS screens to Realtime hub: WebSocket
  • Owner app to API gateway
  • Back office to API gateway
  • Sync service to Order service
  • Realtime hub to Order service
  • API gateway to Menu service
  • API gateway to Inventory service
  • API gateway to Finance & invoicing
  • API gateway to Delivery aggregator: webhooks
  • Order service to Kafka: events
  • Order service to PostgreSQL
  • Inventory service to PostgreSQL
  • Menu service to PostgreSQL
  • Inventory service to Prep planner
  • Inventory service to Variance detection
  • Demand forecast to ClickHouse
  • Prep planner to PostgreSQL
  • Delivery aggregator to HungerStation · Jahez: REST
  • Delivery aggregator to Keeta · ToYou
  • Finance & invoicing to ZATCA Fatoora
  • Order service to Payment provider
  • Order service to WhatsApp Business

Stack

  • Flutter
  • Next.js
  • NestJS
  • PostgreSQL
  • SQLite · offline sync
  • Kafka
  • Redis
  • ClickHouse
  • Python · LightGBM
  • Kubernetes
  • Workflow

A branch day, from forecast to close

  1. 01

    Collect sales history, weather and events

    AutomaticData pipeline

  2. 02

    Forecast demand per item and hour

    AutomaticDemand forecast

  3. 03

    Turn the forecast into prep quantities via recipes

    AutomaticPrep planner

  4. 04

    Approve or adjust the plan

    Branch managerBack office

  5. 05

    Purchase orders and central-kitchen transfers

    PurchasingInventory service

  6. 06

    Take orders in-store and from delivery apps

    Cashier & automaticPOS · Delivery aggregator

  7. 07

    Prepare by station, with a timer per order

    KitchenKDS

  8. 08

    Deduct ingredients and log waste

    Automatic & kitchenInventory service

  9. 09

    Close the day and the branch P&L

    OwnerFinance · Owner app

  • Data & API

The core entities and relations, and a live example from the API.

MenuItem

  • iduuidPK
  • skutextunique
  • name_artext
  • pricenumeric
  • stationenum
  • channel_pricesjsonb
  • is_availablebool

Ingredient

  • iduuidPK
  • nametext
  • unitenum(g,ml,pc)
  • cost_per_unitnumeric
  • shelf_life_hint
  • supplier_iduuidFK

RecipeLine

  • iduuidPK
  • menu_item_iduuidFK
  • ingredient_iduuidFK
  • qtynumeric
  • yield_pctreal

Order

  • iduuidPK
  • branch_iduuidFK
  • channelenum(dine_in,takeaway,hungerstation,jahez,keeta,toyou)
  • external_reftext
  • statusenum
  • totalnumeric
  • placed_attimestamptz
  • ready_attimestamptz

OrderLine

  • iduuidPK
  • order_iduuidFK
  • menu_item_iduuidFK
  • qtyint
  • modifiersjsonb
  • stationenum
  • bumped_attimestamptz

StockMovement

  • idbigserialPK
  • branch_iduuidFK
  • ingredient_iduuidFK
  • kindenum(sale,waste,transfer,purchase,count)
  • qtynumeric
  • unit_costnumeric
  • reasontext
  • attimestamptz

Forecast

  • iduuidPK
  • branch_iduuidFK
  • menu_item_iduuidFK
  • hourtimestamptz
  • qty_p50real
  • qty_p90real
  • model_versiontext

PrepPlan

  • iduuidPK
  • branch_iduuidFK
  • datedate
  • linesjsonb
  • statusenum(draft,approved)
  • approved_byuuidFK
  • MenuItem 1—∞ RecipeLine
  • Ingredient 1—∞ RecipeLine
  • Order 1—∞ OrderLine
  • MenuItem 1—∞ OrderLine
  • Ingredient 1—∞ StockMovement
  • MenuItem 1—∞ Forecast
POST /v1/orders Receive an order from any channel and route it to the kitchen

Request

{
  "branch": "BR-MALQA",
  "channel": "hungerstation",
  "external_ref": "HS-77120458",
  "type": "delivery",
  "items": [
    { "sku": "BRG-CLS", "qty": 3 },
    { "sku": "FRY-L", "qty": 1, "modifiers": ["no_salt"] }
  ],
  "total": 121.00,
  "currency": "SAR",
  "courier_eta": "2026-11-07T21:55:00+03:00"
}

Response

{
  "id": "ord_01M6Q9A4ZK",
  "number": "#4829",
  "status": "accepted",
  "prep_minutes": 14,
  "stations": ["grill", "fryer"],
  "ready_at": "2026-11-07T21:52:00+03:00",
  "einvoice": { "type": "simplified", "status": "queued_for_reporting" }
}

Webhook event · order.ready

{
  "event": "order.ready",
  "order": "#4829",
  "branch": "BR-MALQA",
  "channel": "hungerstation",
  "external_ref": "HS-77120458",
  "prep_seconds": 822,
  "at": "2026-11-07T21:51:40+03:00"
}
  • Integrations

  • HungerStation

    Orders, status updates and menu sync via the partner API

  • Jahez

    Orders, availability and prep times

  • Keeta · ToYou

    Orders and menu sync via partner APIs

  • ZATCA Fatoora

    Stamped simplified invoices with QR, reported within 24 hours

  • Payment provider

    mada and Apple Pay on card terminals paired with the POS

  • WhatsApp Business

    Receipts, loyalty points and offers for opted-in guests

  • Accounting system

    Daily journal postings for sales and cost

  • Security & compliance

  • Data and backups hosted inside the Kingdom.
  • National Cybersecurity Authority Essential Cybersecurity Controls (ECC) across the cloud service.
  • No card data stored: payments run through a PCI DSS-compliant provider and the POS only receives a token.
  • Encrypted on-device database, with remote lock and wipe if a tablet is lost.
  • Role- and branch-based access, with a supervisor PIN for discounts, voids and refunds.
  • Personal Data Protection Law (PDPL) alignment, with explicit opt-in before any marketing message.
  • An audit trail for every void, discount, price change and stock adjustment.
  • Delivery plan

Clear phases with measurable deliverables.

  1. 01

    Discovery & branch visits

    • Observe peak hours in branches and the central kitchen
    • Inventory of menus, recipes and suppliers
    2 weeks
  2. 02

    UX & prototypes

    • POS and KDS prototypes tested with staff
    • Offline sync design
    3 weeks
  3. 03

    POS & kitchen

    • Offline POS and KDS
    • Simplified invoices and payments
    7 weeks
  4. 04

    Delivery & inventory

    • Delivery-app integrations and menu sync
    • Recipes, stock and the central kitchen
    5 weeks
  5. 05

    Forecasting & owner views

    • Train the model on branch sales history
    • Owner dashboard and app
    4 weeks
  6. 06

    Pilot branch & rollout

    • Run one branch alongside the current system, then roll out gradually
    3 weeks
Total duration
24 weeks
Team
Project manager, a business analyst with restaurant-operations experience, 2 backend engineers, a frontend engineer, 2 mobile engineers, a data and ML engineer, and QA.

Performance

  • 24h Offline selling per branch
  • < 5s Delivery order on the kitchen display
  • < 15% Daily forecast error per branch (WAPE)
  • 99.9% Cloud availability

Related work

All work

Have a similar project?

Start a project
Contact us