Kotzilla MCP Server
Overview
The Kotzilla MCP Server connects your AI coding assistant to the Kotzilla Platform. It gives the assistant the same structural and runtime context Kotzilla captures from real user sessions (dependency graphs, resolution timings, crashes, component behavior), together with a knowledge base of proven fix patterns for Android and Kotlin Multiplatform projects.
As a result, the assistant can go from detecting an issue to producing a concrete fix, all inside the editor you already use.
With the MCP Server you can:
- Register your app and configure the SDK in your project through a single natural-language prompt.
- Analyze performance and stability issues from real user sessions across Android and Kotlin Multiplatform targets.
- Get root cause analysis with full dependency resolution trees and timings.
- Get guided fix recommendations and apply them directly in your code.
Key capabilities
Automated setup
Register your app on the Kotzilla Platform and set up the SDK through a single prompt. The assistant reads your project structure (package name, app type), creates the app on the platform, generates kotzilla.json, updates the version catalog and Gradle plugin, and adds the monitoring() call to your Koin setup.
Performance and stability analysis
Analyze performance and stability issues from real user sessions across Android, iOS, Desktop (JVM), JS, and WASM targets. Ask for a lint-style report, compare versions to spot regressions, or rank issues across all your registered apps by severity.
Root cause analysis
Get full dependency resolution trees with timings, identify blocking components on specific screens, trace crashes with symbolicated stack traces, and understand on which thread each event runs.
Guided fixes in your code
Retrieve fix patterns specific to the detected issue type (blocking calls in Koin modules, eager initialization, heavy ViewModel init, StateFlow strategies, constructor work, scope misuse), then let the assistant read your code and apply changes across multiple files under your review.
What you can ask
Once the SDK is set up and your first sessions are recorded, natural language is enough. Here are some examples:
Generate a performance report
Generate a report for my app
Returns a compact lint-style summary with pass/warn/fail status, crash count, ANR count, and prioritized fixes.
Compare versions
Is version 1.2.1 performing better than 1.1.5?
Compares crashes, screen rendering times, issue counts, and component performance between two versions to spot regressions.
Find what to fix first across all your apps
Which issues should I focus on as main priority among my apps?
Generates reports across all your registered apps and ranks them by severity: crashes first, then startup time, ANR, and others.
Root cause analysis
What are the main performance issues in my app?
Lists all detected issues ranked by impact. From there you can ask "Why is it slow?" and the assistant shows what is causing the problem: dependency resolution trees with timings, which components are blocking a specific screen, what is creating ANR risk, and on which thread everything runs.
Fix issues in your code
How can I fix this issue?
The assistant gets fix patterns specific to the issue type: stack traces of crashes, blocking calls in Koin modules, eager initialization, heavy ViewModel init, StateFlow strategies, constructor work, and scope misuse. It reads your code and applies the changes across multiple files under your review.
Investigate crashes
Show me the crashes on the iOS target of my CMP app and help me fix them
Returns symbolicated stack traces with surrounding lifecycle events, and suggests fixes based on the crash context.
Available tools
| Tool | What it does |
|---|---|
list_apps | List your registered apps |
create_app | Register a new app |
guide_sdk_installation | Get SDK setup instructions for your app type |
generate_app_config | Generate kotzilla.json (creates a new API key each call) |
get_platform_activity | See which apps are active and sending data |
get_screen_performance | Screen rendering times and ANR counts |
get_issues | Performance and architecture issues |
get_issue_context | Deep dive: dependency tree, timings, stack traces |
get_performance_guidance | Fix patterns and best practices for a given issue type |
generate_report | Performance summary report |
Authentication and security
On first use, the MCP Server asks you to authenticate with your Kotzilla account. If you do not have an account yet, you can create one for free at console.kotzilla.io.
All communication between the MCP Server and the Kotzilla Platform is encrypted over HTTPS. The same API key scoping and data rules apply as for the SDK and Console.
The MCP Server operates on the same technical, non-sensitive data collected by the Kotzilla SDK and is GDPR-compliant. It does not access any business logic or user-sensitive data from your app.
Why use the MCP Server
- Zero manual onboarding: one prompt registers the app and configures the SDK.
- Context where you work: stay in your editor or terminal while your assistant reasons over real session data.
- Fixes, not just findings: guided remediation combines structural context with proven Kotlin fix patterns.
- Works with any MCP-compatible tool: Claude Code, Cursor, Windsurf, the Android Studio MCP panel, and any LLM-powered terminal assistant.
Get started
To install the MCP Server and register your first app, follow the MCP Server quick start guide.