Skip to content

Vidyax

Code as simple as writing instructions.

Vidyax (.vx) is a programming language designed to make writing code feel like writing plain instructions. Its name comes from the Sanskrit word vidyā (विद्या) — "knowledge".

Vidyax is more than a thin wrapper over Python. It has its own toolchain:

file.vx → Lexer → Parser → AST → Semantic Analysis
        → Compiler → Bytecode (.vxc) → Virtual Machine (C) → Result

Hello world

print "Hello, world!"

vidyax run hello.vx

What makes Vidyax different

  • Four engines, one behavior. A tree-walker, a Python transpiler, a C bytecode VM, and a native compiler run the same program with identical results — guaranteed by differential testing and a fuzzer.
  • Its own Virtual Machine. Written in C, with a mark-sweep garbage collector, an optimizing compiler, bytecode verification, and sandbox limits (instructions, memory, time).
  • Native binaries. vidyax native compiles a program to a standalone executable — no runtime needed to run it.
  • Concurrency. go / wait overlaps slow work (network, AI) with a model that makes data races impossible.
  • Modules & packages. use name splits code across files; vidyax install fetches shared modules.
  • Friendly error messages. Every error is explained in plain language, worded identically across every engine.
  • Built-in AI. An ai module lets you query language models directly from your program.

Where to start

- :material-rocket-launch: **[Guide](GUIDE.md)** — learn Vidyax from scratch through examples; ideal for beginners. - :material-book-open-variant: **[Reference](REFERENCE.md)** — the complete list of syntax, built-in functions, and CLI commands. - :material-chip: **[VVM Specification](VVM_SPEC.md)** — the Virtual Machine architecture, bytecode format, and garbage collector. - :material-arrow-decision: **[Concurrency](CONCURRENCY.md)** — the design of `go` / `wait` and how tasks work across all four engines.

Running through the Virtual Machine

# Compile to bytecode
vidyax bytecode program.vx

# Run with the C VM
./vm/vxvm program.vxc

The output is identical to vidyax run program.vx, but it runs on a self-built C runtime — not on top of Python.

Built by NaDev · from a student's hands, for students.