Hi HN, I'm the creator of Flint.
Over the years, I got tired of CI/CD pipelines breaking because a 300-line Bash script had a silent error, or waiting for Python to boot and load modules just to parse a JSON payload.
Flint is a strictly typed, compiled language built specifically for infrastructure scripting. It uses an arena allocator (zero GC) and compiles to standalone native binaries. It's not general-purpose (no async, no floats, no web servers). It's built for one thing: replacing fragile glue code with fast, safe native binaries.
Happy to answer any questions about the compiler architecture or our zero-copy I/O implementation!
I often wonder how these types of projects compare to nushell. I love nushell because of its use of pipes. It doesn't compile to native code but it is my goto for this sort of thing.
"Wait, isn't this just like Nushell?"
I get this question a lot. I love Nushell for interactive terminal work. But Flint is fundamentally different. Nushell is an interactive shell environment; Flint is an AOT-compiled systems language.
If you want to explore JSON files interactively in your terminal, use Nushell or jq. But if you need to write a bulletproof CI/CD script, compile it into a tiny 300KB native binary, and drop it on a production server without installing ANY runtime or dependencies... that's when you use Flint.
Flint is a strictly typed, compiled language built specifically for infrastructure scripting. It uses an arena allocator (zero GC) and compiles to standalone native binaries. It's not general-purpose (no async, no floats, no web servers). It's built for one thing: replacing fragile glue code with fast, safe native binaries.
Happy to answer any questions about the compiler architecture or our zero-copy I/O implementation!