01
Mechanized Comprehension
How large-scale software systems can be understood through tooling that extracts meaning, structure, and intent from code at scale.
Legacy Code Labs · Est. 2024 · Research since 2019
Advancing the art and science of AI-era software.
Tooling, experiments, and research to systematically understand and evolve codebases in the AI-era.
01 Four pillars of research.
The work is organized around four themes that emerged from years of building tools and applying them in real-world engagements.
02
Mechanized Verification
How correctness and gaps in evolving software can be identified systematically through autonomous analysis.
03
Mechanized Remediation
How problems, once identified, can be addressed through targeted, tool-assisted intervention rather than costly rewrites.
04
Directed Evolution
How codebases can be intentionally steered toward better states through continuous, guided transformation, not accidental drift.
02 Open source.
Open source tools born from real engagements. Each one built to solve a problem that manual approaches couldn't.
Clarity
A software design tool for AI-native developers and coding agents. Review design impact before commit.
Verification
Eureka
Visualizes the structure of large Kotlin and Java classes as interactive graphs. A 3,000-line class on one screen, updating in real-time.
Comprehension
Cardbox
Structural analysis of Android codebases through graph queries via jQAssistant.
Comprehension
Timelapse
Leverages Git history to reveal tribal knowledge. Every codebase has a story in its version control. Timelapse makes it legible.
Comprehension
03 Origin.
February 20, 2017: I picked up Michael Feathers' Working Effectively with Legacy Code while struggling with a codebase I couldn't make sense of. By 2019, I was building tools to solve it. What once took nine months to earn, a team's trust, collapsed to two or three days.
Read the full story04 Publications.
Papers, tools, and open source releases from the lab.
2025VerificationClarity: software design for AI-native developers and coding agents2023ComprehensionEureka: breakthrough tooling for Kotlin and Java codebases2021ComprehensionCardbox: structural analysis of Android codebases via jQAssistant2019ComprehensionTimelapse: leveraging Git history for codebase comprehension
-Additional research being curated from the archives.
Feathers wrote the book that named the problem.
We're building the tools that solve it.