AML (AI Machine Language) is a modern programming language designed to bridge the gap between natural language and high-performance C++ code. It leverages large language models to translate human-readable instructions into optimized C++, which is then compiled to native binaries using industry-standard tools.

The AML project aims to democratize systems programming by allowing developers to express their intent in plain English, while still producing code that runs as fast as hand-optimized C++. This approach combines the productivity of high-level languages with the performance of low-level systems programming.

Key design principles of AML include:

• Natural Language First: Write code in plain English, let AML handle the translation to efficient C++.
• Performance by Default: Generated code is compiled with G++, ensuring native execution speed.
• Simplicity: No complex syntax to learn—just describe what you want in clear language.
• Extensibility: Use any LLM backend (OpenAI, LM Studio, local models) for code generation.

• AI-Powered Code Generation: Convert natural language to C++ using state-of-the-art LLMs.
• Native Performance: Generated code compiled with G++ for maximum speed.
• Flexible LLM Backend: Support for OpenAI, LM Studio, and custom API endpoints.
• Zero Runtime Overhead: No interpreter or VM—just pure compiled binaries.
• Human-Readable Output: Generated C++ code is clean, commented, and ready for review.
• Simple Configuration: Easy setup with model name, API key, and endpoint URL.

This tutorial will guide you through installing and using AML to create your first program.

Configure your preferred LLM backend:

For local development with LM Studio:

Create a file called hello.aml:

Output:

• OpenAI (GPT-4, GPT-3.5)
• LM Studio (local)
• Anthropic Claude
• Google Gemini
• Any OpenAI-compatible API

• ./aml config set model [your model]
• ./aml config set apikey [your api key]
• ./aml config set posturl [chat complation url]
• ./aml config show [posturl | apikey | model]
• ./aml config unset [posturl | apikey | model]
• ./aml compile [your source code (test.aml)] [output binary (test)]
• ./aml version

Q: Do I need an API key?
A: Yes for cloud LLMs, no for local setups like LM Studio.

Q: What compilers are supported?
A: G++ are officially supported.

Q: Where can I download old versions?
A: Link.

Q: Is AML free?
A: Yes.