Smeron
All Work
Secure communication

SLC

  • UI/UX case study
  • Real-time architecture
  • Hybrid serverless
SLC — Secure communication

A communication suite with secure real-time messaging, high-fidelity video conferencing and VOIP-to-PSTN world calls.

Who it was built for

The business, and what they came with.

Built for

SLC

A communication suite for privacy-conscious users and small teams. Messaging and video are free; calling real phone numbers anywhere in the world is the paid part.

The brief

Chat and video are table stakes and cost money to run, so the business rests on World Calls — VOIP out to landlines and mobiles. That means credit accounting has to be exact and calls have to connect on the first try.

What we delivered

What SLC has now that they did not have before.

  • Messaging that feels instant

    Realtime Database drives delivery, presence and typing indicators, so conversation keeps the rhythm people expect from the app they left.

  • Video and voice that hold up

    GetStream and Twilio handle adaptive bitrate video and conference audio, including on the connections users actually have.

  • Credit accounting that cannot drift

    Cloud Functions initiate each PSTN call and deduct credit atomically. In a prepaid product a billing bug is not a bug, it is a refund queue and a review problem.

  • Notifications that arrive

    A dual path through Expo and direct FCM, because a message the app never surfaced is indistinguishable from a broken product.

Where the growth comes from

Why each piece of the build pays for itself.

  • The paid feature reaches the real world

    Anyone can ship free chat. Dialling an actual phone number is what gives a messaging app a revenue line.

  • Credits, not seats

    People top up when they travel or take on international work, so spending scales with their need rather than a plan they have to justify.

  • Delivery is retention

    Push reliability decides whether the app is checked daily or eventually deleted, which decides whether credits get bought at all.

The business analysis

Market, model and architecture. Every figure is either cited or labelled as a model.

SLC gives away the two features people compare messaging apps on and charges for the one nobody advertises: dialling an actual telephone number. That inversion is the whole business, and it puts an unusual amount of commercial weight on plumbing. Free chat and video are a cost centre that buys distribution. World Calls is the revenue line, and it fails commercially not when a call sounds bad but when the credit ledger disagrees with the user about what they were charged.

Wholesale voice carrier market, 2025
$40.26bn
Forecast to $68.61bn by 2030 at 11.25% CAGR. This is the market SLC buys termination from, not the one it sells into.

Source · Mordor Intelligence, Wholesale Voice Carrier Market

Share of wholesale voice carried over IP
71.3%
2024 share. The reason a small app can offer international termination at all: the underlying network is already IP.

Source · Mordor Intelligence, Wholesale Voice Carrier Market

Of the VoIP services market is international long distance
60.1%
2024 share, with mobile VoIP international services growing fastest. SLC's paid feature sits in the largest slice.

Source · Mordor Intelligence, VoIP Services Market

Independent push delivery paths
2
Expo and direct FCM. A message the app never surfaced is indistinguishable from a broken product.

Model · Counted from the delivered scope in this case study.

The paid feature sits in the one part of telecoms still growing

Messaging is a commodity and video conferencing has been free since 2020. What is not free is terminating a call onto the public telephone network in another country, and that market is neither small nor shrinking. Mordor Intelligence puts the wholesale voice carrier market at $40.26bn in 2025, rising to $68.61bn by 2030 at an 11.25% compound rate, with VoIP already carrying 71.3% of wholesale voice in 2024 1. International long distance accounted for 60.1% of the VoIP services market in the same year 2.

Wholesale voice carrier market, USD billions

The supply side of SLC's paid feature. The 2030 column is the research house's forecast and is drawn hollow.

  • $40.3bn
  • $68.6bn
  • 2025
  • 2030

Hollow columns are the published forecast, not a measurement.

Source · Mordor Intelligence, Wholesale Voice Carrier Market

Why the billing code is the most important code in the product

In a prepaid product, a billing defect is not a bug. It is a refund queue, a chargeback, and a one-star review that mentions the word "stolen". Cloud Functions initiate each PSTN call and deduct credit atomically, which is the correct architecture for exactly one reason: the deduction and the call must not be able to disagree.

Failure modeWhat the user experiencesCommercial consequence
Credit deducted, call never connectedMoney taken for nothingRefund, and a review that costs more than the refund
Call connected, credit not deductedA free callSilent margin leak, discovered at reconciliation
Deduction applied twiceBalance drains faster than expectedThe most damaging of the three, because it looks deliberate
Push notification not deliveredA message that never arrivedNot a billing fault, but it is why the app gets deleted before credit is ever bought

Where the engineering risk sits, weighted by commercial damage

Our weighting of the four delivered capabilities by how expensive their failure is, rather than by how hard they were to build.

Credit accounting and call initiation40%
The only component whose failure produces a refund and a public complaint.
Push delivery25%
Determines whether the app is opened daily, which determines whether credits are ever purchased.
Real-time messaging20%
Sets whether the app feels like the one the user left. Retention, not revenue.
Video and conference audio15%
Expected, expensive, and rarely the reason anyone switches apps.

Model · Our weighting of the four capabilities documented in this case study by the cost of their failure. Judgement, not measurement.

One thread, three modes

The product's design argument is that chat, video and a real phone call are the same conversation at different intensities. Commercially, that is a funnel: the free mode is where the user lives, and the paid mode is one tap away at the moment they need it.

From a free message to a billed minute

The escalation path is the monetisation path. The offshoots are the states a prepaid calling product must handle without argument.

  1. Encrypted chat

    Delivery, presence and typing state over Realtime Database, so conversation keeps the rhythm people expect.

  2. Escalate in place

    The same thread becomes HD video or conference audio. No new room, no link, no second app.

  3. Dial a real numberdecision

    Balance checked, call initiated by a Cloud Function, credit deducted atomically as it connects.

    • ↳ Insufficient credit → top-up prompted before dialling, never after
    • ↳ Termination fails → no deduction, and the user is told why
  4. Top up

    Credits added in-app at the moment of need, which is the only moment a prepaid user willingly pays.

  5. Notified, and still installed

    Delivery through two independent push paths, because an app that misses messages is deleted long before it is monetised.

A hybrid serverless split, and why it is drawn this way

Media rendering and real-time state sit on the client; anything that touches money sits in functions the client cannot reach around.

  1. Client

    Carries media rendering and real-time UI state. Fast, and deliberately not trusted with anything financial.

    • React Native 0.81
    • Expo 54
    • TypeScript
    • Zustand
  2. Real-time data

    Message delivery, presence and typing indicators, where latency is felt directly by the user.

    • Firebase Realtime Database
  3. Trusted server logic

    Call initiation, credit deduction and notification fan-out. The atomic operations live here because the client must not be able to authorise a call.

    • Firebase Cloud Functions
    • Firestore
  4. Media and telephony

    The parts it would be reckless to build: adaptive video, conference audio and PSTN termination.

    • GetStream Video SDK
    • Twilio Voice
    • Twilio Video
  5. Identity and delivery

    Who the user is, and making sure the app can reach them at all.

    • Firebase Auth
    • FCM
    • Expo push

What we would watch

RiskWhy it bitesEarly indicator
Termination rate exposureGross margin per minute is set by wholesale rates SLC does not control, and rates move by destinationMargin per minute drifting on specific country prefixes rather than overall
Fraud on prepaid balancesCredit-based calling is a standard target for traffic pumping and stolen-card top-ups; both convert directly into lossTop-up volume concentrated on a small number of expensive destinations
Free-tier cost outrunning top-upsMessaging and video cost money per active user whether or not they ever call a phone numberInfrastructure cost per active user rising while top-up frequency is flat

References

  1. 1.Wholesale Voice Carrier Market — size, share and trends · Mordor Intelligence
  2. 2.VoIP Services Market — size, share and growth trends · Mordor Intelligence

One tap between three modes

The path a real user takes, in order.

  1. Chat

    Encrypted messaging with presence and typing state.

  2. Escalate

    The same thread becomes an HD video or audio call.

  3. Dial out

    Call a landline or mobile anywhere, billed against credits.

  4. Top up

    Credits added in-app, deducted per call as it connects.

Who it is for

The people whose problem this solves, and what they came for.

  • Privacy-conscious users

    Want an encrypted alternative to the mainstream apps.

  • Travellers and expats

    Need cheap, reliable calls to ordinary phone numbers.

  • Small teams

    Want a private, branded channel without enterprise licensing.

Reference

For the technical reader. Everyone else has what they need above.

Recommended case studies

The closest work to SLC, scored for relevance rather than picked by position.

Want the version of this built for your business?

Fifteen minutes with the people who shipped it. No deck, no account manager.