🧬 RepoDNA v1.4.4 and v1.4.5 Are Now Stable!
I've reached another development milestone with RepoDNA, my open-source, local-first repository intelligence project.
Over these two releases, I've been focusing on building a stronger native evidence-explanation and ranking foundation.
Here's what has improved:
✅ Offline native evidence explanations
✅ Optional local evidence-ranker training
✅ Reproducible training and stronger data validation
✅ New evaluation and calibration diagnostics
✅ Better findings exports and filtering
✅ Expanded security and regression testing
A particularly important part of this work has been making evaluation more transparent rather than simply introducing new functionality.
The aim is to keep improving RepoDNA as a practical tool for developers who want to understand their repositories without unnecessary reliance on external services.
RepoDNA v1.4.5 is available now:
https://github.com/sanskarIN/RepoDNA/releases/tag/v1.4.5
You can also explore the browser application:
https://sanskarin.github.io/RepoDNA/
Thank you to everyone following the development of my open-source projects.
Your feedback, encouragement, and optional support help keep independent development moving forward.
More projects: https://sanskarin.github.io/
Understand your codebase. See its DNA.