docs / guide / The Cajeta Guide

00 — Introduction

Cajeta is a compiled language with Java-shaped syntax, C++-grade native output, and Rust-style memory safety without a garbage collector. Programs compile through LLVM to real binaries. Memory is managed by ownership: = borrows, # transfers, and the compiler proves the rest.

A first taste:

package hello;

import cajeta.lang.System;

public class Main {
    public static void main(String[] args) {
        System.stdout.println("hello, cajeta");
    }
}

What comes with the toolchain:

• One binary, cajeta — compiler, build tool, package manager, doc generator, and debug server.
• A standard library covering collections, streams, fibers and async, file and network I/O, time, hashing, codecs, math, and GPU kernels.
• Compile-time dependency injection and aspects — no runtime container.
• A capability model: a program touches only what its manifest declares.
• Built-in support for AI agents: cajeta compiler-mcp is a Model Context Protocol server inside the compiler, serving skills — hand-written implementation guidance for the language, the toolchain, and every library. More than 180 ship in the binary, and every published library carries its own (chapter 03).

How the documentation fits together

• This guide — a linear walk: install the toolchain, build a project, then the language and standard library in order. Read it front to back.
• Standard library reference — one page per class, organized by package. Go here for full method surfaces.
• The tour — runnable demos of every feature. Chapters link to the tour file that demonstrates what they teach.
• Language specification — design-level reference.

Next: Installation.

Source: docs/guide/00-introduction.md · 1 min · 236 words