The GitStat Stack: What We Use & Why
Building software is mostly about making decisions. And usually, you retain the right to regret those decisions six months later.
With GitStat, we wanted to move fast. Like, "idea to production in a weekend" fast. But we also care about aesthetics and performance, which usually fight against speed.
So here’s the stack we picked to keep us sane and shipping.
The Heavy Lifters
Next.js 16 + React 19
We're living on the bleeding edge. Is it risky? Maybe. Is it fun? Absolutely.
We use the App Router because organizing files by routes just makes sense to our brains. Server Components handle the heavy lifting of fetching data, keeping the client bundle small. React 19 brings the new exciting stuff, and honestly, we just like new shiny toys.
Bun
Waiting for npm install is so 2022. Bun is fast. Ridiculously fast. It handles our package management and local dev server without breaking a sweat. It just works, and it respects our time.
Making It Pretty
Tailwind CSS v4
We upgraded to v4 alpha/beta because we hate configuration files. The new engine is instant, and defining variables in CSS instead of a tailwind.config.js feels like the way it should have always been.
If you see us centering a div, know that we're doing it with utility classes and zero shame.
Radix UI + Motion
Building accessible components from scratch is a noble cause, but we have features to ship. Radix primitives handle the keyboard navigation and focus management so we don't have to.
Then we sprinkle motion (formerly Framer Motion) on top because things that pop in and slide out just feel better. The human brain likes things that move nicely. We obey the brain.
The Data Stuff
Supabase
Managing a Postgres instance ourselves? No thanks.
Supabase handles our database and authentication. The "Sign in with GitHub" flow was basically free. It lets us focus on the app logic instead of wondering if our database backups are actually running.
Octokit + DataBuddy
Since we're analyzing GitHub repositories, we talk to the GitHub API a lot. Octokit is the standard for that.
DataBuddy helps us manage some of the more complex data operations without reinventing the wheel.
Visualizing The Numbers
Recharts
Charts are hard. Recharts makes them less hard. It's composable, React-native, and lets us customize the look until it fits our "clean aesthetic" requirements.
We need those star history graphs to look smooth, not like a jagged mess from a 1998 Excel spreadsheet.
Why this stack?
We didn't pick these tools because they're trendy (okay, maybe a little bit). We picked them because they remove friction.
We don't want to spend days configuring Webpack. We don't want to write boilerplate SQL. We want to write code that actually does something for you.
This stack stays out of our way. And that's the best kind of stack.