Dev.to AI 🤖 Ai 👁 0 📖 4 min read

Leksika: Building a Programming Language. with Russian Syntax — Architecture & 2026 Roadmap

What if programming did not require English keywords? Leksika is an open-source, Russian-first programming language built around Russian words, morphology and grammatical cases. Instead of translating familiar English k

What if programming did not require English keywords?

Leksika is an open-source, Russian-first programming language built around Russian words, morphology and grammatical cases. Instead of translating familiar English keywords, the project explores a different idea: let the grammar of the language itself shape the way code is written.

The project is developed under the GitHub account @zmlsaggg.

A programming language that speaks Russian

Leksika is designed so that keywords, variable names and core language constructs use Russian.

A small example:

пусть сумма равно пяти.
положить десять в сумму.
прибавить два к сумме.
сказать речь Сумма равна (сумма) конец речи.

The expected output is:

Сумма равна 12

One of the central ideas is grammatical case. The form of a word can express its role in an operation:

  • в сумму — a destination;
  • к сумме — the value being increased;
  • сумму — the value being operated on.

The project also includes morphology for Russian word forms, number words, collections, conditional statements and loops.

This is an experiment in language design, not simply a conventional programming language with translated keywords.

Try it without installing anything

The online playground lets you experiment with Leksika directly in a browser.

It includes a code editor, examples, execution output and an AI assistant. The assistant has two separate modes:

Code mode focuses on programming. It can review syntax, explain problems and propose a corrected version of a program. Suggested changes are not silently inserted: you decide whether to apply them.

Chat mode is a general conversational assistant. It can discuss programming, explain concepts and answer questions beyond the language itself.

The AI layer supports two approaches:

  • Run compatible public Hugging Face models in the browser using WebGPU or WebAssembly, depending on device support.
  • Connect an external OpenAI-compatible API, Anthropic or Gemini by providing your own endpoint, model and key.

There is an important privacy distinction: a local browser model processes your code in the browser, while an external API provider receives the prompt and code you send to it. API keys are not stored in local storage.

The playground is an educational browser environment, not a complete replacement for the desktop runtime.

Leksus: the engine behind the language

The runtime and development tools are called Leksus.

The repository contains the language implementation, examples, documentation, tests, installers and the evolving native engine.

The current desktop runtime still uses Python 3.8+ as a transition layer. This is stated openly in the project documentation. The long-term goal is to remove that dependency from normal execution.

At the same time, a native execution path is being developed:

  1. Read a UTF-8 .lek source file.
  2. Parse supported language constructs.
  3. Compile them into Leksika's own bytecode.
  4. Verify instructions, registers and constants.
  5. Execute the bytecode in a native virtual machine.

The native core uses its own value model and exact decimal arithmetic for the supported subset. It is currently built through a temporary C11 bootstrap. The project is not yet fully self-hosting, and the native compiler does not support every feature of the language.

That distinction matters: the goal is not just to make the existing implementation faster, but to build an execution environment whose semantics belong to Leksika itself.

Technical documentation:

What makes the project interesting?

There are several connected research and engineering directions:

1. Russian morphology in code

The project explores how declension, grammatical case, gender, number and word forms can participate in programming syntax.

2. A language-specific virtual machine

Leksika is moving toward its own bytecode, value semantics, compiler and runtime instead of permanently relying on another language's execution model.

3. A large Russian word-form index

The project builds a binary index from a pinned, MIT-licensed Russian word-form source. The index is used by the native engine and is also being connected to browser-based word suggestions.

4. A browser playground with AI assistance

People can try the language without installing the desktop runtime, ask questions and get help understanding or correcting code.

5. An open development process

The repository includes tests, examples, technical plans and documentation. The implementation status is documented rather than presenting unfinished components as completed features.

Who might find Leksika useful?

Leksika may be interesting to:

  • Developers experimenting with programming-language design and compiler architecture.
  • People interested in Russian NLP, morphology and computational linguistics.
  • Students and educators looking for a different way to introduce programming concepts.
  • Open-source contributors interested in virtual machines, parsers and language tooling.
  • Anyone curious about what a programming language designed around Russian grammar could look like.

It is not intended to replace Python, C++, Rust or other established languages today. It is an independent language project exploring a different design direction.

How to explore or contribute

The easiest way to start is to open the browser playground:

https://leksika.gleeze.com/playground.html

If you want to inspect the implementation, read the source code:

https://github.com/zmlsaggg/Leksika

You can also browse the latest builds, read the contribution guide, or explore the examples directory.

Contributions around grammar, morphology, tests, documentation, compiler design and native runtime development are welcome.

Project links

Leksika is an open-source experiment in Russian-first programming, grammar-aware syntax and an independent language runtime.

If that sounds interesting, try writing your first program in Russian.

📰 Read the original article on Dev.to AI

Originally published by Dev.to AI. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.