How to Develop a Mobile App for Self-Service Laundry?
Coins and queues at the terminal — the main pain point for self-service laundry owners. The customer spends 5 minutes looking for change, another 2 choosing a program through a murky screen. With the app: scan the QR on the machine, choose a program, pay, get notified when ready. For the network owner — remote machine monitoring, load statistics, dynamic pricing without visiting the location. We have been developing mobile apps for self-service laundry for over 5 years — tested on 12+ networks, processed over 1 million wash cycles, and connected over 800 machines. The average cost per wash cycle is around 300 rubles, and the app pays for itself in 6–8 months by increasing machine load by 25%.
MQTT is the key protocol for communication with machines. It provides latency under 100 ms, 10 times faster than HTTP polling. Architecture: IoT module ↔ MQTT broker ↔ backend ↔ mobile app via WebSocket.
How MQTT Ensures Communication with Machines?
Washing machines in self-service are controlled via an IoT module, built-in or installed parallel: ESP32 or Raspberry Pi with GSM/Wi-Fi. The module connects to the machine's control board via relays (button emulation) or via UART/RS485 if the machine has a service interface. Most commercial machine manufacturers (Electrolux Professional, Miele Professional, Speed Queen) provide an API or at least a service protocol description — you need to request directly from the vendor. Cheap machines without a protocol are controlled via relays: the module sees a "cycle started" signal from a current sensor (SCT-013) and sends the state to the server.
MQTT with a broker (Mosquitto/HiveMQ) ensures a persistent connection without HTTP overhead. Each machine publishes its status to the topic laundry/{id}/status — the app subscribes and receives updates instantly. Connection loss is compensated by Last Will Testament. Protocol comparison:
| Protocol | Latency | Server Load | Module Power Consumption |
|---|---|---|---|
| MQTT | <100 ms | Low | Low |
| HTTP polling | 1-30 s | High | High |
| WebSocket | <50 ms | Medium | Medium |
// Android: subscribing to machine status via MQTT class LaundryMachineMonitor(private val machineId: String) { private val mqttClient: MqttAndroidClient = /* initialization */ fun subscribeToMachine(onUpdate: (MachineStatus) -> Unit) { mqttClient.subscribe("laundry/$machineId/status", 1) { _, message -> val json = String(message.payload) val status = Json.decodeFromString<MachineStatus>(json) onUpdate(status) } } fun startCycle(program: WashProgram, token: String) { val command = Json.encodeToString(StartCycleCommand(program, token)) mqttClient.publish("laundry/$machineId/command", command.toByteArray(), 1, false) } } @Serializable data class MachineStatus( val state: MachineState, // IDLE, RUNNING, DONE, ERROR val programName: String?, val remainingSeconds: Int?, val errorCode: String? ) How to Bypass App Store Commission When Paying for Laundry?
Key issue: Apple considers topping up an in-app wallet a "digital good" and requires IAP with a 30% commission. But if the wallet is used to pay for physical services (laundry is a physical service), you can use external acquiring directly. Scheme: top up balance → redirect to Safari/SafariViewController with a web payment page (YooKassa, Stripe, CloudPayments). Payment for a specific cycle — deduction from balance via API. Apple Guidelines 3.1.5(b) allows this for "real goods and services." On Android with Google Pay it's easier: PaymentsClient with a card or integration in WebView. Savings on commission — up to 30% per transaction.
Why Should Booking Be Paid?
The user wants to know if a machine is free before going to the laundromat. A map of locations with real-time machine availability indicators is the main function of the home screen. Filtering: "only with free machines," "with dryers."
Booking a machine for 10–15 minutes is a controversial function. Without booking: you arrive and all are occupied. With booking: many abandoned reservations. Compromise: paid booking (1 conditional unit deducted), credited toward the cycle. Our experience: implementing paid booking reduced empty reservations by 70%.
Push notification 5 minutes before cycle end and upon completion — via FCM/APNs. On the server side: a worker checks the remaining time based on machine data, schedules a push via FCM Schedule (Android) or APNs with apns-expiration.
Accumulating loyalty points per wash cycle — a simple retention mechanic. Every Nth cycle free. Implementation on the server, mobile app shows progress via API.
What's Included in the Project
- API documentation for IoT module and mobile app integration.
- Mobile app source code (iOS/Android) in Swift 5.9+ / Kotlin with Jetpack Compose.
- Server module in Node.js or Python with MQTT broker.
- Integration guide for any machines: API, UART/RS485, relays.
- Staff training on managing the network via CMS.
- Technical support for 3 months after release.
Process Workflow
| Stage | Duration | What We Do |
|---|---|---|
| Analytics | 1-2 weeks | Study machine fleet, choose protocol, audit current business processes |
| Design | 2-3 weeks | UX design (map, booking, payment), IoT network architecture |
| Development | 4-8 weeks | Module firmware, mobile code, server, admin panel |
| Testing | 2-3 weeks | Integration testing with real machines, MQTT load testing |
| Deployment | 1-2 weeks | Install modules in laundries, publish to App Store / Google Play |
Checklist for Machine Integration
- Determine connection type: does the machine have a service API or only relays?
- For API: request documentation from the manufacturer, check MQTT support.
- For relays: select current sensor SCT-013 and ESP32 module with Wi-Fi.
- Set up MQTT broker (Mosquitto) on the server.
- Test start and stop cycle commands manually.
- Integrate payment gateway (Stripe, YooKassa) for balance top-up.
- Set up push notifications via FCM/APNs.
- Test booking: set a timer for 10-15 minutes with fund hold.
Indicative Timelines
- MVP (one laundry, basic functionality): 6–8 weeks.
- Full solution with map of locations, loyalty program, and CMS: 4–5 months.
Cost is calculated individually — depends on the number of machines, integration complexity, and need for App Store approval. Our team's experience (10+ years in mobile development, Apple and Google certifications) reduces risks and timelines. Order development today and get a presentation with examples of completed projects. Contact us for a consultation — we will evaluate your project for free.