Metor is a fully autonomous, high-performance programming language and AI tensor engine built completely from scratch using Rust and LLVM. It features its own native toolchain (Forge), a built-in package manager (Metos), high-concurrency socket networking, and a custom VS Code extension identity.
- Dedicated File Extension (
.mtr): Clean, custom syntax extension completely independent from other ecosystems. - LLVM-Backed Compiler: Engineered in Rust, utilizing a custom lexer, parser, type checker, and LLVM IR code generator to output native machine code (
.o). - Metos Package Manager: Full project bootstrapping (
forge new), dependency configuration (metos.toml), and automatic module linking frommetos_modules/. - High-Performance Runtime & AI Engine:
- Native HTTP Server (
serve): Multi-threaded TCP/HTTP server loop built directly into the core runtime. - Matrix Multiplication (
matmul): Low-level hardware-accelerated tensor operations for AI and machine learning workloads. - Zero-Dependency JSON Parser (
json_get): Fast payload extraction for web APIs.
- Native HTTP Server (
- Custom VS Code Integration: First-class syntax highlighting and a custom branded icon theme (
.mtrfile badges).
Metor/
├── editors/
│ └── vscode-metor/ # VS Code extension, syntax grammar, and icon themes
├── src/
│ ├── cli/ # Forge command-line interface and arguments parser
│ ├── compiler/ # Lexer, Parser, Type Checker, Codegen, and Linker
│ └── main.rs # Compiler entry point
├── Cargo.toml # Rust compiler workspace manifest
└── README.md
Getting Started
Prerequisites
Rust & Cargo (latest stable edition)
LLVM / Inkwell dependencies
C Compiler (cc / clang / gcc) for native linking
1. Build the Forge Toolchain
Clone the repository and compile the compiler using Cargo:
Bash
git clone [https://github.com/adi6206096675/Metor.git](https://github.com/adi6206096675/Metor.git)
cd Metor
cargo build --release
2. Create a New Metor Project
Initialize a new project using the Forge CLI:
Bash
cargo run -- new my_backend
cd my_backend
This generates a structured workspace containing:
metos.toml — Package configuration and dependency declarations.
src/main.mtr — Main source entry point.
3. Install Dependencies
Resolve and fetch package modules into metos_modules/:
Bash
cargo run --manifest-path ../Cargo.toml -- install
4. Build or Run Your Program
Compile your .mtr source file into a standalone native executable:
Bash
cargo run --manifest-path ../Cargo.toml -- run src/main.mtr
📝 Example Metor Code
Code snippet
// Simple HTTP Server and Tensor Processing example in Metor
let msg = "Hello from Metor AI Engine!"
print msg
fn handle_req() {
send_http("HTTP/1.1 200 OK\r\nContent-Type: text/plain\r\n\r\nHello from Metor Server!")
}
// Start the native server loop on port 8080
serve(8080)
🛡️ License
Distributed under the MIT License. See LICENSE for more information.
Build under AROM Labs