Source: handbook/ in the repository. Every code block compiles: CI runs ray check on each one.

The raylang handbook

Guides for building real things with raylang, from ray new to the installed app or the deployed service. Each guide is a complete path for one kind of project: it does not walk the language module by module (the manual and the reference do that), but shows which pieces to use, in what order and why.

The guides build on each other: the notes app that starts as a mobile app comes back as a desktop app, as an API, as a server-rendered site and as a site with a React frontend. Each one stores its data differently, so by the end you have used the whole range: the std/kv store, SQLite, Postgres, files and Redis.

All the code on these pages compiles: CI runs ray check on every block.

Chapters

Getting started: install, the editor, a project, the language in fifteen minutes, concurrency, working with an LLM assistant and the tools.
Mobile app for iOS and Android: a notes app with a React + TypeScript frontend and data in std/kv, from the project to the simulator, the emulator and hot reload on the phone.
Cross-platform app: the same app on macOS, Linux and Windows, with SQLite, native menus and dialogs, and an interface that adapts to every screen.
Windows in depth: everything std/ui gives a desktop app, on a text editor: the bridge to the page, menus, dialogs, window kinds and unsaved changes.
Server-rendered site: the notes as a server-rendered website, with compiled templates, forms, Markdown and the data in files.
Web API: the notes as a JSON API with the web framework, Postgres with a connection pool, token authentication and graceful shutdown.
Site with a React frontend: a React frontend embedded in the binary, a JSON API and the notes in Redis.
Command-line tool: a command-line tool with options, piped input, output for people and for programs, and exit codes.
LLMs and MCP: an assistant that writes raylang verified with ray mcp, an agent written in raylang with the llm, mcp and agent packages, and an MCP server of your own.
Performance: measuring, profiling with ray profile, fixing the algorithm, compiling to native and spreading work across cores, with measured figures.
Shipping: packaging and signing for each system, publishing to the mobile stores and signed automatic updates on the desktop.
More examples: the apps of the ray-language organization, by use case, for what the chapters do not cover.

For any function, the reference has its signature, and the manual (Spanish) explains the language in depth.