zig run versus zig build-exe

Zig gives you two main paths from source to result: zig run for quick iteration, and zig build-exe when you want a standalone binary to keep.

The command zig run compiles your file and immediately executes it, all in one step. Nothing is left behind on disk afterward.

Because it skips saving a file, zig run is ideal while you are still experimenting. Change code, run again, and see output in seconds.

When you want an executable you can ship or run later, use zig build-exe. It compiles your source into a real file on disk.

By default the output is named after your file, like main, and placed in the current directory. You then run it yourself.

You can choose the binary name with the -femit-bin flag. This is handy when one source file should produce a specifically named tool.

Even zig run builds a binary behind the scenes and caches it. A second run with no changes reuses that cache and starts almost instantly.

With zig run, anything after -- is forwarded to your program as its command-line arguments rather than to the compiler.

zig build-exe is a direct compiler command for a single artifact. For full projects you will later graduate to zig build with a build.zig file.

Ask one question: do I need to keep this binary? If no, reach for zig run. If yes, reach for zig build-exe and grab the output file.

Both commands use the very same Zig compiler, so the code that runs is identical. Only the convenience and the leftover file differ.

You are testing a tiny script and do not need to keep any file around. Which command fits best?

You now have two tools: zig run for fast compile-and-go iteration, and zig build-exe to produce a real binary you can name, keep, and ship. 🚀