BLE Connected-Device Platform for a D2C Wellness Brand
How we built and operate the full app, backend, and BLE stack behind 40K+ connected devices
Executive Summary
App downloads with 10-15K monthly active users
Connected BLE devices in the field
Retention, with a crash rate under 1%
From kickoff to first app-store release with a 2-3 person team
The Challenge
A D2C consumer wellness hardware brand needed to launch BLE-connected devices with a companion software ecosystem, but it had no in-house app engineering team. The app would become the primary control surface for the hardware: it needed to connect reliably to BLE devices, support low-latency control, manage content, enable remote partner features, collect product analytics, and remain stable enough to protect app-store reviews and customer trust. This was a greenfield build — no prior companion app existed. The client had BLE hardware, a Shopify D2C storefront, Klaviyo email, and CleverTap analytics, and the ambition to launch a connected-device product without building an internal engineering organization. Without software, the hardware risked becoming a commodity product; differentiation depended entirely on the quality of the mobile apps, backend, realtime layer, and device protocol.
Key Pain Points
- ✕No in-house mobile, backend, infrastructure, BLE, or realtime engineering capacity
- ✕Hardware at risk of becoming a commodity without a software experience
- ✕App stability directly tied to app-store reviews and customer trust
- ✕No protocol authority to hold hardware manufacturers accountable
Our Solution
Amasa built and maintains a complete 5-repo platform: native Android app, native iOS app, backend, admin backoffice, and infrastructure-as-config — going from kickoff to first app-store release in about 5 weeks with a 2-3 person team. The Android app is Kotlin with MVVM, Dagger 2, the Nordic BLE library, and ExoPlayer; the iOS app is Swift with MVVM-C, Combine, and native CoreBluetooth with background BLE mode. The backend runs Express.js and TypeScript in Docker on EC2 with MongoDB Atlas, AWS Cognito JWT authentication, and a dedicated ELK monitoring stack. The realtime layer uses an in-process Aedes MQTT broker with JWT-authenticated connections. The platform controls hardware through 8-byte command frames written to a GATT characteristic for lowest latency, supports 11 vibration modes, long-distance partner control pairing two users through a 6-digit code with ~60fps touch streaming over MQTT, and custom patterns recorded as 15ms-step time series with a public gallery, trending, and creator analytics. Amasa also authored the formal BLE requirements specification used to source and qualify chip manufacturers — making the apps the source of protocol truth. Amasa continues to operate as the client's full app and platform engineering team on a maintenance retainer with a 24-hour SLA.
Implementation Approach
Platform Architecture
Designed the 5-repo ecosystem: native iOS and Android apps, backend, admin backoffice, and infrastructure-as-config
BLE Protocol & Apps
Built the 8-byte GATT command protocol, 11 vibration modes, and native Kotlin and Swift apps with background BLE
Realtime & Content Layer
Shipped MQTT-based partner control at ~60fps, custom pattern recording, and a public pattern gallery
Operations & Protocol Governance
Authored the formal BLE spec for manufacturer qualification and continue running the platform under a 24-hour SLA retainer
The Results
The implementation delivered transformative results across all key metrics, with immediate impact on operational efficiency, accuracy, and customer satisfaction.
Impact Metrics
| Metric | Before | After | Improvement |
|---|---|---|---|
| Devices in field | Greenfield | 40K+ connected devices | From zero |
| User scale | No prior app platform | 100K+ downloads, 10-15K MAU | From zero |
| Stability | Greenfield | >90% retention, <1% crash rate | Store-review safe |
| Engineering capacity | No internal software team | Amasa team of 2-3 across the full stack | Full coverage |
Key Takeaways
- A hardware brand can launch and operate a connected product without hiring an internal engineering organization
- Owning the BLE protocol specification turns hardware suppliers into interchangeable vendors who must conform to the app
- An in-process MQTT broker delivers internet-scale realtime control of physical hardware without separate infrastructure
- Retention above 90% and crash rate under 1% are what protect app-store ratings for hardware companion apps
Inside the Engagement
The Starting Point
This was a greenfield build. No prior companion app existed. The client had BLE hardware, a Shopify D2C storefront, Klaviyo email, CleverTap analytics, and the ambition to launch a connected-device product without building an internal engineering organization. The main constraint was not replacing a manual workflow – it was launching a software-dependent hardware business with no mobile, backend, infrastructure, BLE, or realtime engineering capacity in-house.
The Platform
- Android: Kotlin, MVVM, Dagger 2, Retrofit and OkHttp, auto token refresh, Nordic BLE library, LiveData, StateFlow, coroutines, ExoPlayer, Android API 24+.
- iOS: Swift, MVVM-C, Combine, Alamofire, native CoreBluetooth with background BLE mode, Valet keychain, iOS 14+.
- Backend: Express.js and TypeScript in Docker on EC2 (ap-south-1), Nginx SSL termination, node-cron, MongoDB Atlas via TypeORM, AWS Cognito JWT auth, Firebase, Google Sheets, Klaviyo, CleverTap, and a dedicated ELK monitoring stack.
- Realtime: an in-process Aedes MQTT broker in the same Node process as the API; mobile clients connect via the Gojek Courier MQTT SDK with JWT-authenticated connections over TCP and WebSocket.
Flagship Capabilities
The platform controls BLE hardware through 8-byte command frames written to a GATT characteristic using Write Without Response for lowest latency, with 11 vibration modes and scan filtering by advertising UUID.
Long-distance partner control pairs two users through a 6-digit code. One user’s trackpad touch position is sampled at around 60fps, streamed over MQTT, and converted into BLE motor commands on the partner’s device – real internet-scale remote control of physical hardware.
Custom patterns let users record touch intensity as a 15ms-step time series, save it, replay it, loop it, and publish it to a public gallery with trending, leaderboards, and creator analytics.
Amasa also authored the formal BLE requirements specification used to source and qualify chip manufacturers – covering advertising, the GATT database, the 8-byte control protocol, the Jieli OTA service, and acceptance tests. The apps became the source of protocol truth.
Full Results
| Metric | Before | After |
|---|---|---|
| Devices in field | Greenfield | 40K+ connected devices |
| User scale | No prior app platform | 100K+ downloads and 10–15K MAU |
| Retention | Greenfield | >90% retention |
| Crash rate | Greenfield | <1% crash rate |
| Support model | No internal app engineering | Ongoing maintenance retainer with a 24-hour SLA |
| Engineering team | No app, backend, or infrastructure team | Amasa team of 2–3 covering iOS, Android, backend, backoffice, and infra |
The Impact
The client launched and operates a connected-product business without hiring an internal app or backend engineering team. The hardware became an app-driven experience with partner remote control, custom patterns, content, analytics, and realtime behavior. The BLE protocol spec became a strategic asset: hardware vendors must conform to the app, giving the brand manufacturer flexibility without rewriting the software layer.
The Takeaway
Amasa built and operates the iOS, Android, backend, realtime, backoffice, cloud, and BLE protocol stack for a D2C connected-device brand – scaling to 100K+ downloads, 40K+ connected devices, 10–15K MAU, >90% retention, and <1% crash rate while acting as the brand’s full software engineering team.
Quick Facts
Industry
Consumer Hardware
Solution Type
IoT & Connected Devices
Published
August 24, 2026
Technologies Used
Related Resources
Ready to Achieve Similar Results?
Let's discuss how we can transform your business with AI solutions tailored to your specific needs