A self-hosted, dynamically typed programming language that compiles directly to native Windows x64 and Linux x64 executables.
Self-hosted compiler · Reference compiler · Latest release · Language guide
MiniLang emits native PE32+ and ELF64 images from its own x86-64 backends. The production compiler is written in MiniLang and reaches a byte-identical self-hosting fixed point. A separately maintained Python reference compiler produces compatible executables and provides a transparent bootstrap path.
function fibonacci(n)
if n < 2 then return n end if
return fibonacci(n - 1) + fibonacci(n - 2)
end function
print fibonacci(10)The language and standard library include structured error handling, garbage collection, threads and synchronization, networking and TLS, FFI, conditional compilation, documentation comments, formatting, and a native unit-test framework.
| Project | What it demonstrates |
|---|---|
| MiniLangCompilerML | The self-hosted MiniLang compiler for Windows and Linux x64 |
| MiniLangCompilerPy | The Python reference and bootstrap compiler |
| MiniDoom | A full DOOM engine port with classic and OpenGL renderers |
| MiniQuake | A playable Quake source port preserving Protocol 15 |
| MiniSQL | A transactional SQL server with concurrency, TLS, and replication |
| MiniDoc | A self-hosted documentation generator and source analyzer |
More projects explore Quake II, 2D games, native GUIs, installers, and a lightweight MiniLang IDE. Browse the complete repository collection.
The two compiler implementations are continuously checked against shared language behavior and native output. Self-hosted stage builds, Windows/Linux targets, object-pipeline builds, and representative applications are measured for correctness, binary compatibility, build time, and memory use. The detailed evidence and current measurements live in the compiler repositories instead of being reduced to unsupported headline claims.
MiniLang and its project family were developed with extensive assistance from generative AI. The repositories keep the implementation, tests, generated API documentation, performance reports, and limitations visible so the results can be inspected and reproduced.
Questions, bug reports, and focused contributions are welcome in the relevant project repository.
