Workshops: Agentic Coding
Agentic Coding for iOS & Android Development: Workshop
Mobile development brings its own hurdles for AI agents: Xcode and Gradle builds, simulators, code signing and platform guidelines. This workshop shows iOS and Android teams how to plug agents into these toolchains and evolve native and cross-platform apps faster and safely.
Who is this workshop for?
- iOS developers (Swift, SwiftUI)
- Android developers (Kotlin, Jetpack Compose)
- Cross-platform teams using React Native, Expo or Flutter
- Mobile tech leads
- Teams turning web apps into native apps
What you will take away
- Have agents run builds via xcodebuild and Gradle and fix errors on their own.
- Feed the agent results from simulators and emulators via screenshots and UI tests.
- Implement SwiftUI and Compose screens from design specs in a structured way.
- Use agents for platform tasks such as permissions, localisation and privacy declarations.
- Know which steps, like signing and store upload, should deliberately stay with humans.
Agenda
Module 1
Opening mobile toolchains to agents
- Claude Code, Cursor, GitHub Copilot and OpenAI Codex alongside Xcode and Android Studio
- Defining build and test commands for the agent
- Project rules for architecture (MVVM, clean architecture) and modules
Module 2
UI with agents
- Exercise: build a SwiftUI or Compose screen from a template
- Simulator and emulator screenshots as feedback
- Dark mode, dynamic type and accessibility
Module 3
Cross-platform
- React Native and Expo, Flutter: what agents do well
- Native modules and bridges
- Reusing web code: Capacitor or a WebView approach like WebViewGold
Module 4
Tests and quality
- Unit and UI tests with XCTest, Espresso or Maestro
- Having crash logs analysed
- Performance and app size
Module 5
All the way to the store
- Preparing permission texts, privacy labels and the Data safety form
- Release notes and localisation
- Keeping signing, CI and approvals deliberately manual
Why mobile is different
An agent in the terminal does not automatically see what happens in the simulator, and long builds slow every feedback loop. Without adaptation, agents quickly lose value in mobile projects.
We show how to prepare build scripts, test targets and visual feedback so agents work efficiently, and which tasks, such as signing certificates, do not belong in an agent's hands.
Grounded in app practice
neooutput is a venture by jocapps GmbH, which has been building apps since 2014; apps from this background reach more than 45 million end users. The ecosystem includes WebViewGold and appsubmitter.io for store publishing.
- Experience with native and cross-platform apps
- Knowledge of typical store review hurdles
- Pragmatic choice between native, cross-platform and WebView
Format
Hands-on with your app project, remotely or on-site, in English or German. For iOS exercises, participants need a Mac with Xcode.
Frequently asked questions
Can the agent work directly in Xcode?
Agents like Claude Code work in the terminal alongside Xcode and use xcodebuild for builds and tests. Xcode also has its own AI features; we show how both fit together.
Is this worth it for Flutter or React Native?
Very much so. Cross-platform code is often especially accessible to agents. We also cover the native parts where agents tend to struggle.
Will the agent handle store publishing too?
It can prepare texts, privacy declarations and release notes. We recommend keeping signing, upload and release manual or in controlled CI pipelines.
Workshop
Get in touch
Want agents to make your app team noticeably faster?
Tell us a little about your team – we will get back to you within 24 hours with a proposal for content, format and dates.
Prefer to talk?
More workshops
- Workshops: Agentic CodingAgentic Coding for QA & Test Automation: Workshop
- Workshops: Agentic CodingAgentic Coding for Data Science & Analytics: Workshop
- Workshops: Agentic CodingAgentic Coding for DevOps & Infrastructure as Code: Workshop
- Deploy & App StoresHow to Publish a Lovable App on Google Play (Step-by-Step Guide)
