Maero Lab
MaeroLab is a macOS application and private, on-device AI tool that maps your codebase so you can understand and debug complex software faster.
Why did I build this app?
Managing hundreds of Xcode projects (my other side quests) makes tracing functions and revisiting complex architectures time-consuming. I built this to quickly visualize structure and understand codebases at a glance. And I also wish this application was available when I first started to learn iOS swift development.
Is it solving a real problem?
It is solving an emerging problem. As AI-assisted coding accelerates, developers are increasingly working with large codebases they didn’t write line by line, especially in collaborative environments where multiple engineers contribute, making it difficult to understand what someone else built. Even small changes can require tracing hundreds of lines of logic. It is designed for this shift: a private, on-device tool that helps you understand, navigate, and debug complex AI-generated applications with clarity.
Managing hundreds of Xcode projects (my other side quests) makes tracing functions and revisiting complex architectures time-consuming. I built this to quickly visualize structure and understand codebases at a glance. And I also wish this application was available when I first started to learn iOS swift development.
Is it solving a real problem?
It is solving an emerging problem. As AI-assisted coding accelerates, developers are increasingly working with large codebases they didn’t write line by line, especially in collaborative environments where multiple engineers contribute, making it difficult to understand what someone else built. Even small changes can require tracing hundreds of lines of logic. It is designed for this shift: a private, on-device tool that helps you understand, navigate, and debug complex AI-generated applications with clarity.
The Features
1. Visualize Entire Codebase: Drag and drop your project into an interactive map to see files, symbols, and relationships all in one place. It helps you quickly understand structure and spot impact areas before making changes.
2. Walkthrough: Generates a guided explanation of your project architecture and how key pieces fit together. It breaks complex systems into clear steps so you can onboard faster and reason about code with confidence.
3. Chat: Ask architecture and behavior questions about a selected file and get context-aware answers grounded in your project map. It helps you learn faster, debug intent, and understand how one file affects others.
4. Code Autopsy: Highlight specific lines and get a precise, plain-language explanation of what the code is doing. It is ideal for decoding unfamiliar logic, validating assumptions, and reducing reverse-engineering time.
5. What If: Analyzes proposed changes before you implement them so you can evaluate risk early. It highlights likely impact zones, dependencies, and tradeoffs to support safer refactors.
6. Connected Files: Displays how files and components relate through incoming and outgoing links. It helps you trace dependency chains and understand ripple effects across the codebase.
7. Offline & Privacy: All of the analysis runs on-device, so MaeroLab works even without internet access. Your code stays local to your machine, giving you a privacy-first workflow by default. Powered by Apple Foundation Models.
Limitation
Apple Foundation Models: Apple Foundation Models are optimized for on-device privacy, speed, and efficiency, not maximum reasoning power. Compared with larger cloud “frontier” LLMs, they have a smaller context/token budget, so performance can drop on very large codebases, long multi-step reasoning, and highly complex prompts.
However, Apple Foundation Models are continuing to improve and as they mature, this application will become significantly more capable over time.
Next steps
MaeroLab is built for where software development is heading, not just where it is today. As AI systems become more powerful, software can also become harder to understand and debug. Large cloud LLMs are incredibly useful, but many rely on data collection and infrastructure that are not always transparent to users. I wanted to build a more private alternative so people can create and debug confidently while keeping their code in their own control.
Download & try it out →
Download & try it out →
Role: iOS Engineer, UX/UI Designer
Time: 2 months (Concept originated while building Banau)
Softwares: Swift, SwiftUI, Apple Foundation Models (LLM), Codex, Claude Code, Anti-gravity, Perplexity
AI Integration Strategy: AI was used to support high-level architecture and system design. Not to generate the product step by step. With Swift’s evolving syntax, increasing UI complexity, and updates to the Observable framework, blindly prompting implementation details would inevitably introduce inconsistencies and technical debt, regardless of which model is considered “the best.”
Time: 2 months (Concept originated while building Banau)
Softwares: Swift, SwiftUI, Apple Foundation Models (LLM), Codex, Claude Code, Anti-gravity, Perplexity
AI Integration Strategy: AI was used to support high-level architecture and system design. Not to generate the product step by step. With Swift’s evolving syntax, increasing UI complexity, and updates to the Observable framework, blindly prompting implementation details would inevitably introduce inconsistencies and technical debt, regardless of which model is considered “the best.”