Logo
LaunchPad Next

Overview - AI Workspace


Launchpad Next includes an AI Workspace — a set of documentation files that describe your project to AI tools before they generate code. Instead of re-explaining your design system, folder structure, and conventions on every prompt, that information lives in one place and gets read automatically.

The result is AI-generated code that's more accurate, consistent, and easier to maintain.

Why It Exists

Without project context, AI tends to:

  • Create duplicate components
  • Ignore existing design patterns
  • Use inconsistent styling
  • Place files in the wrong locations
  • Introduce patterns that don't fit the codebase

The AI Workspace solves this by giving every prompt a shared foundation.

Workspace Structure

All context files live inside ai/:

ai/
├── PROJECT_CONTEXT.md
├── DESIGN_SYSTEM.md
├── FILE_STRUCTURE.md
├── COMPONENT_RULES.md
├── SEO_ARCHITECTURE.md
├── PROMPT_TEMPLATE.md
├── DECISIONS.md
├── CHANGELOG.md
└── TASKS.md

What Each File Does

| File | Question it answers |
|---|---|
| `PROJECT_CONTEXT.md` | What is this project? |
| `DESIGN_SYSTEM.md` | How should it look? |
| `FILE_STRUCTURE.md` | Where should new files go? |
| `COMPONENT_RULES.md` | How should components be built? |
| `SEO_ARCHITECTURE.md` | How does SEO work? |
| `PROMPT_TEMPLATE.md` | How should AI work inside this project? |
| `DECISIONS.md` | Why was it built this way? |
| `CHANGELOG.md` | What changed recently? |
| `TASKS.md` | What comes next? |

PROJECT_CONTEXT.md

High-level overview of the project. Covers goals, tech stack, design philosophy, development standards, and architecture. This is the first file any AI tool should read.

DESIGN_SYSTEM.md

Visual design rules: colors, typography, spacing, border radius, shadows, animation, buttons, and icons. Keeps generated UI consistent with the rest of the project.

FILE_STRUCTURE.md

Folder organization, naming conventions, and file placement for components, content, and SEO files. Prevents AI from creating files in the wrong location.

COMPONENT_RULES.md

How components should be structured — reusability standards, TypeScript conventions, content separation, and rules for using Motion and Button components.

SEO_ARCHITECTURE.md

Metadata architecture, structured data, Open Graph rules, and dynamic SEO patterns. Ensures generated pages follow the same SEO approach as the rest of the project.

PROMPT_TEMPLATE.md

A reusable workflow for AI-assisted development: recommended prompting patterns, project instructions, change logging, and a development checklist.

DECISIONS.md

A log of architectural decisions and the reasoning behind them. Helps AI avoid re-litigating settled choices.

// Examples
Use Motion instead of Framer Motion
Store SEO inside page content files
Use content-driven architecture

CHANGELOG.md

A running record of what has changed. Helps AI understand recent modifications before making new ones.

// Examples
Added PricingTable component
Updated Hero layout
Improved mobile navigation

TASKS.md

Backlog and future work. Useful context for planning and prioritization prompts.

// Examples
Add integrations page
Improve blog layout
Create dashboard templates

How AI Uses These Files

Before generating code, AI should read:

ai/PROJECT_CONTEXT.md
ai/DESIGN_SYSTEM.md
ai/FILE_STRUCTURE.md
ai/COMPONENT_RULES.md
ai/SEO_ARCHITECTURE.md
ai/DECISIONS.md
ai/CHANGELOG.md

This gives the AI a complete picture of the project's architecture, visual language, coding conventions, past decisions, and recent changes — before writing a single line.

Example Workflow

Say you prompt:

Create a new pricing section with yearly billing support.

A context-aware AI would:

  1. Read the design system for colors, spacing, and animation rules
  2. Check existing pricing components to avoid duplication
  3. Follow component rules for structure and TypeScript standards
  4. Reuse the existing Button component
  5. Apply Motion animations consistently
  6. Place the file according to the file structure guide
  7. Log the change in CHANGELOG.md

Without the AI Workspace, steps 1–7 become guesswork.

Recommended Prompt

Start every session with:

Read the files inside ai/ before making changes.
 
Follow the project's design system, component rules,
file structure, and SEO architecture.

Many AI tools (Cursor, Claude Code, GitHub Copilot) support repository-level instruction files that inject this automatically.

Repository AI Files

Launchpad Next includes top-level instruction files for popular AI tools:

AGENTS.md
CLAUDE.md
.github/copilot-instructions.md

These files let AI tools discover project rules without requiring manual prompting on every session.

Supported Tools

The AI Workspace works with any context-aware AI tool:

  • Claude Code
  • Cursor
  • GitHub Copilot
  • ChatGPT
  • Gemini CLI
  • Windsurf
  • Aider
  • OpenCode
  • Continue

Keeping the Workspace Current

The workspace is only as useful as it is accurate. Update it whenever:

  • Something changes → log it in CHANGELOG.md (files modified, features added, bugs fixed)
  • A decision is made → record it in DECISIONS.md with the reasoning
  • Standards evolve → update the relevant documentation file

Stale documentation produces the same inconsistent output as no documentation.

Benefits

| Without AI Workspace | With AI Workspace |
|---|---|
| AI guesses at conventions | AI follows your actual conventions |
| Duplicate components appear | Existing components get reused |
| Inconsistent styling | Design system is applied correctly |
| Wrong file locations | Files placed correctly first time |
| Repeated explanations per prompt | Context loaded once, reused everywhere |

Summary

The AI Workspace is your project's memory. It gives every AI tool a shared understanding of what exists, how it works, and why it was built that way — before any code is written.

ai/
├── PROJECT_CONTEXT.md   ← What is this project?
├── DESIGN_SYSTEM.md     ← How should it look?
├── FILE_STRUCTURE.md    ← Where do files go?
├── COMPONENT_RULES.md   ← How are components built?
├── SEO_ARCHITECTURE.md  ← How does SEO work?
├── PROMPT_TEMPLATE.md   ← How should AI work here?
├── DECISIONS.md         ← Why was it built this way?
├── CHANGELOG.md         ← What changed recently?
└── TASKS.md             ← What comes next?

Keep it current, reference it in every prompt, and your AI-generated code will consistently reflect the project it belongs to.

NextPrompt Engineering