Menu

LanguageEnglish
Talk to us

Structure, components, testing, release

You can check the results at the end of each stage.

Progress

What you get at each stage

  1. Structure

    Data and permissions defined before screens

    Data model · Permission table · Screen list

  2. Components

    Screens from CoUI; login, payments, and notifications from existing feature code

    First screens

  3. Testing

    Automated tests, then approval by the engineer in charge

    Test report · Approval record

  4. Release

    Web and app released from the same code, store submission included

    Store build · Web address

We don't build from scratch every time

We reuse the features every service needs and spend the time on what is unique to yours.

  • Login

    Login · Access

    Social login, role-based access

  • Pay

    Payment · Payout

    Card payments, subscriptions, refunds

  • Notify

    Notifications

    Push, KakaoTalk alerts, email

  • Admin

    Admin

    List search, filters, permissions, activity logs

AI writes the code, people make the decisions

AI agents write the first draft of the code; engineers decide what to build and what to ship.

AI agent

What AI does

  • Drafting code from the spec
  • Writing test code
  • Assembling repeated screens
  • Summarizing changes

Engineer

What people do

  • Designing data and access
  • Reviewing and fixing code
  • Checking payments and access
  • Approving releases

Tests in five layers

From a single calculation at the base to scenarios on cloud devices at the top, each layer automatically checks something different.

  1. UNIT

    Calculations and rules

    Rules like price calculation and permission checks, tested function by function

    Every change

  2. WIDGET

    Screen behavior

    Whether the screen responds as intended when you tap and type

    Every change

  3. GOLDEN

    Screenshot comparison

    Screens compared as images per device and theme, shared as a gallery

    Every change

  4. CATALOG

    Screen catalog

    Every screen and component, collected by state; a missing screen blocks the merge

    Every change

  5. BDD · DEVICE

    Scenarios on cloud devices

    Usage scenarios written as sentences, run on Android and iOS devices in AWS and Google Cloud

    Nightly

Test results feed into the checks before release

Three checks before release

Only work that passes all three goes on to engineer review.

  1. Matches the agreed scope

    Behavior outside the agreed scope fails

  2. Feature and screen tests

    From single functions up to full screens

  3. Impact on other features

    Checks that a fix hasn't broken other features

When all three pass, it goes to engineer review

What you'll receive in the first week

In the consultation, we show you example deliverables for each stage, based on the service you want to build.