Let's be honest: stateless AI code assistants are incredibly powerful, but they have terrible short-term memory. They look at your code through a keyhole. I built CXGRD to give AI the full macro-perspective of your codebase so it stops breaking things.
A few months ago, I was working on a PR reviewer tool built specifically for solo devs and small teams (that tool failed, btw). Every single time I asked an AI assistant to fix one thing or implement a new feature, it would inadvertently break code somewhere else. It made me wonder: Why can't AI tools have a full map of the entire project directory? Dependencies in complex systems are highly intertwined, and stateless AI models desperately need that macro-context to know exactly which files to modify. I needed a way to give these models the precise context required for every new change. So, I opened my editor and built a CLI tool to do exactly that. That's how CXGRD was born.
Since I work on this product alone, I don't have to answer to a board of directors or chase artificial hyper-growth. That means I get to focus entirely on making the tool better for you. Here is what I promise:
Hey, I'm Manan
I'm a self-taught programmer who loves to dive into the complex architecture of large systems and explore how things work under the hood. I get a massive kick out of "builder's dopamine"—there's nothing quite like the feeling of shipping a clean feature and seeing it solve a real-world problem. When I'm not actively pushing commits or optimizing infrastructure for CXGRD, you can usually find me reading heavy technical books, messing around with low-level systems, or thinking about my next side project.
CXGRD is shaped entirely by the people who use it. If you have a feature request, found a bug, or just want to talk shop about code, my inbox is always open. You can follow my building journey on X (Twitter) or check out the project over on GitHub