Hey PH π
We built CodeBoarding because AI coding is getting fast, but understanding what changed is getting harder.
Itβs open-source, and it makes a kind of mind map for your codebase.
Instead of just looking at folders or asking an LLM to index everything, we use static analysis / control-flow graphs to find the actual architecture, then use a small LLM layer to name the clusters so humans can read it.
Would love feedback from people building with coding agents.
CodeBoarding is a recursive visualiser for large codebases. It generates accurate representations of your code by using static analysis and LLMs. It starts from the highest level of abstractions and lets you dive deeper recursively.