NUASM is a universal assembler that lets you write low‑level code in 51 human languages — including Hindi, Japanese, Spanish, Arabic, Quechua, Basque, Swahili, and many more.
No transpilation.
No DSL tricks.
No “English‑only mnemonics”.
Just real assembly, expressed in your own language.
NUASM is built in pure Python and outputs real x86‑64 machine code. It includes:
51 language packs (with dialects and regional variants)
Kids Mode for teaching low‑level concepts
Localized error messages
A universal tokenizer and encoder
Support for bootloaders, kernels, and low‑level systems
A fully documented wiki with examples in every language
The goal is simple:
If you can speak it, you can program it.
Repo: https://github.com/cyberenigma-lgtm/NeuroUniversalASM (github.com in Bing)
Happy to answer questions, discuss design decisions, or help anyone build new language packs.
Assembly mnemonics were invented in the 1950s by English‑speaking engineers, and the entire world inherited that constraint. NUASM removes it. The assembler doesn’t care which human language you use — it maps everything to the same machine code.
This project is not about replacing English. It’s about removing barriers.
If you’re curious about:
how the tokenizer handles multilingual input
how mnemonics are mapped across languages
how dialects are normalized
how Kids Mode works
or how to add your own language pack
I’m here and happy to dive deep.