Skip to content

baltazarparra/guia

Repository files navigation

guia

A restrained, bilingual landing page that explains how to start building software with AI coding agents.

guia is not a tool or a framework. It is a public guide plus a set of agent-friendly artifacts that walk the reader through a structured development workflow, from choosing a stack to writing a PRD to executing through pre-implementation plans.

Built with Vite, React, Tailwind CSS, React Three Fiber, and GSAP. Deployed as a static site on GitHub Pages.

Live site: baltazarparra.github.io/guia

Why this exists

Most people starting with AI coding agents jump straight into prompting and hope for the best. The result is often fast chaos instead of fast delivery.

guia presents a simple, practical workflow that helps beginners build with more structure. It covers stack decisions, planning documents, execution models, reusable templates, and direct entrypoints for code agents.

Stack

Layer Technology
Build Vite
UI React (JavaScript)
Styling Tailwind CSS
3D React Three Fiber + Drei
Animation GSAP + ScrollTrigger
State bridge Zustand
Deployment GitHub Actions + GitHub Pages

For code agents

Use case What to read
Bootstrap a new project llms.txt — paste this URL into your agent. It will interview the user and create PLAN.md, IMPLEMENTATION-ROADMAP.md, and AGENTS.md.
Need a static overview start-here.md — lightweight guide for humans or agents that only got the home URL or cannot render the SPA.
Need a machine index agent-index.json — machine-readable map of public entrypoints, templates, examples, and repo docs.
Need filled examples examples/README.md — sample outputs of PLAN.md, roadmap, and AGENTS.md.
Contribute to guia AGENTS.md — project rules, constraints, and validation workflow.

Local development

Prerequisites: Node.js 18+

npm install
npm run dev

Build for production:

npm run build
npm run preview

Validate code quality:

npm run check    # lint + format check + typecheck + tests
npm run lint:fix # auto-fix lint issues
npm run format   # auto-format all files

Project structure

src/
  components/
    layout/       PageShell, SectionContainer, LanguageToggle
    sections/     HeroSection, GlossarySection, AgentsSection, ModesSection,
                  ModelsSection, PlanSection, RoadmapSection, ExecutionSection,
                  BootstrapSection, TemplatesSection, ClosingSection, DipCtaSection
    ui/           Button, SectionHeading, CopyBlock, Modal, Accordion
    three/        NotebookScene, DiamondModel, SceneController, SceneLights,
                  WebGLErrorBoundary
  content/        Bilingual copy (pt.js, en.js), diamondStates.js, references.js
  hooks/          Scroll progress, language, reduced complexity, scene state
  styles/         Global styles, Tailwind extensions
roadmap/          IMPLEMENTATION-ROADMAP.md (task-level execution guide)
examples/         Filled example outputs generated with the guia workflow
templates/       Reusable templates for agentic development
  README.md      Index and usage order
  PLAN-TEMPLATE.md, ROADMAP-TEMPLATE.md, RULES-TEMPLATE.md,
  PRD-TEMPLATE.md, SKILLS-TEMPLATE.md
docs/             Internal PRDs, copy outlines
public/           Static assets and agent entrypoints (favicon, OG image, llms.txt, start-here.md, agent-index.json)
.github/
  workflows/      deploy.yml (GitHub Pages)

Architecture

The page is a single scroll experience with two parallel layers:

  • DOM layer — content sections scroll normally, animated with GSAP ScrollTrigger
  • 3D layer — a fixed <Canvas> behind the DOM renders the diamond, driven by scroll progress via Zustand

GSAP never animates Three.js objects directly. R3F never reads scroll position from the DOM. A Zustand store bridges the two systems.

The 3D diamond is a progressive enhancement. The page is fully functional without WebGL.

Languages

Content is available in PT-BR and EN, switchable via a toggle that preserves scroll position.

Documentation

Document Purpose
PLAN.md Full product spec — vision, stack, architecture, phases, risks, acceptance criteria
IMPLEMENTATION-ROADMAP.md Task-level execution guide — 57 atomic tasks across 7 phases
AGENTS.md Instructions for AI coding agents working on this project
examples/README.md Filled examples that show the expected output shape

Contributing

If you'd like to contribute, start by reading PLAN.md for context and IMPLEMENTATION-ROADMAP.md for the current execution plan.

License

MIT

Releases

No releases published

Packages

 
 
 

Contributors