Ziran

A general-purpose language · Experimental

Start with the language.
Build what matters.

Ziran turns checked .zi modules into native code or portable programs. Write a command-line tool, a library, a service, or an interface without making the language depend on a UI framework.

Write your first program See what works today

Open source Checked modules Honest target coverage

hello.zisource
Score :: struct {
    value: s32;
    bonus: s32;
}

Total :: (score: Score) -> s32 {
    return score.value + score.bonus;
}

Answer :: () -> s32 {
    score: Score = Score.{
        value = 40,
        bonus = 2
    };
    return Total(score);
}
one source.zirC · C++ · Go · .zib
01 / WRITEReadable modules

Records, enums, functions, imports, generics, fixed arrays, and borrowed slices in the tested subset.

02 / CHECKShared semantics

The checker validates source and saved IR before native generation or portable execution.

03 / RUNChoose a target

Generate C, C++, or native Go. Bundle supported programs as .zib.

The pipeline

One language.
Explicit edges.

Ziran keeps the language, its checked intermediate form, and its portable runtime in one repository. A target that cannot support a feature must report it instead of silently changing the program.

Read the architecture
SOURCE.ziModules and imports
CHECKED FORM.zirSaved, versioned IR
NATIVE OUTPUTC · C++ · Go
PORTABLE BUNDLE.zib

Made to compose

Ziran is the language.
Kryon is a library.

Kryon’s UI modules are written in Ziran and imported like any other library. Widgets, layout, styling, and rendering belong to Kryon; they are not special language syntax. A Ziran program with no Kryon import needs no graphics runtime.

The two-repository migration is still underway. Kryon’s checked modules build, while platform integration and downstream application cutover need more work.

Explore Kryon

Current state

Useful now.
Still experimental.

The compiler and portable runner support a tested subset; neither full Jai syntax nor complete backend parity is finished. Owned growable collections, general law proofs, automatic parallel execution, and GPU execution remain planned work.

Read the implementation status