Macroscope is built on a core perception layer that processes and synthesizes code activity across the entire product development workflow. The key part of this perception layer is our code walking system. Code walkers traverse the Abstract Syntax Tree (AST) of the code to build a graph-based representation of the entire codebase.
This approach allows Macroscope to:
- Understand large, complex, multi-repository codebases.
- See precisely how each code change impacts your product.
- Detect bugs that most other code review tools often miss, because those tools have a shallower understanding of your code.
- Generate technical summaries that reflect the true state of your product.
FAQ: What languages does Macroscope code review support?→ Macroscope code review works with every file type in your pull request. We have native code walkers for Golang, Python, TypeScript, Java, Rust, Kotlin, Swift, Ruby, and Elixir, which provide the deepest AST-based analysis. All other files — configuration, documentation, scripts, and more — are also reviewed.
FAQ: What models does Macroscope use?→ Macroscope doesn’t rely on a single model. Our Auto-Tune system automatically tests multiple model, prompt, and parameter combinations to find the best-performing configuration for each language. An LLM-driven curator analyzes failures and proposes improvements each iteration, so the pipeline continuously gets better as new models are released. Learn more about Auto-Tune on our blog. 
