Modern Kotlin apps, whether they target Android or are built with Kotlin Multiplatform (KMP) and Compose Multiplatform (CMP) are composed of dozens (or hundreds) of modules. Each one contains components such as ViewModels, Repositories, Services, and screens that interact across feature boundaries, scopes, threads, and now platforms.
As your architecture scales, so does the risk of hidden performance issues, crashes, or subtle bugs that only surface on specific devices or targets in production.
The Kotzilla Platform is built to give Kotlin developers using the Koin framework a clear, high-fidelity view of what's really happening in your apps at runtime, across every supported target, without requiring access to your implementation logic or user data.
In this article, we'll break down exactly what the platform collects, how it handles data, and why your code remains private and secure whether you're debugging locally, in production, or through the Kotzilla MCP Server from your AI coding assistant.
The Kotzilla Platform is composed of four components:
The SDK collects the data that helps you detect issues like:
The SDK works in the background without requiring manual traces or code modifications. It operates in both development and production environments and is tightly integrated with Koin, using its container as a signal source.
The Kotzilla SDK collects technical metadata about how your components and dependencies behave at runtime. This includes:
Component & dependency structure
Lifecycle & performance metrics
Crash data
Version-aware metrics
This gives your team precise visibility into the runtime characteristics of your app, across every target, without requiring manual tracing.
To be explicit, here’s what the Kotzilla SDK does not collect:
| ✅ Collected | ❌ Not Collected |
|---|---|
| Class/type metadata | Method bodies or source code |
| Dependency resolution stats | Business logic or internal algorithms |
| Component lifecycle events | UI screen contents, input field data |
| Thread and performance signals | Network request or response payloads |
| Crash type, stack frames, frame origins | Screenshots, video recordings |
| Device type, OS version, app version | Personally Identifiable Information (PII) |
In release builds, all class and method names are obfuscated by tools like R8 or ProGuard. To map the obfuscated names back to the original ones in the Console, you upload a mapping file. Mapping files are managed by the Kotzilla Gradle plugin and uploaded only if you opt in. You stay in full control of what is shared.
On iOS, the same principle applies. Crashes are captured on-device with PLCrashReporter and stored as raw addresses by default. To get readable Kotlin stack traces in the Console, you upload your dSYM file with the setupKotzillaXcode script. As with Android mapping files, dSYM upload is opt-in and entirely under your control.
The Kotzilla Platform meets modern security and data protection requirements:
The Kotzilla MCP Server connects the Kotzilla Platform to AI coding assistants (Claude, Cursor, and others). It allows developers to query their app's performance data and get AI-assisted remediation guidance without leaving their IDE.
The MCP Server exposes three categories of tools:
For the complete and current list of tools, see the Kotzilla MCP Server reference.
The Kotzilla Platform gives you deep runtime visibility into how your Kotlin app's components behave, across every Android and Kotlin Multiplatform target, without compromising code privacy.
It captures technical metadata, not business logic, to identify issues related to component or dependency resolution delays, thread misuse, slow screens and crashes. In production, class and method names are obfuscated by default, and mapping file and dSYM upload are entirely under your control.
The Kotzilla MCP Server brings the same insights into your AI coding assistant, with org-scoped access, no AI training on customer data, and no source code or PII in transit.
With secure, encrypted communication and zero user data collection, the platform is built to support high-trust environments from early development to post-release debugging and AI-assisted remediation.
If you're maintaining a large Kotlin or Kotlin Multiplatform codebase, you are using Koin, and you need actionable insights without exposing sensitive logic, the Kotzilla Platform is built for that job.