BYU Strategy - Marriott School of Business

AI Tool Reference

Quick Reference

This page provides a practical reference for AI-assisted development tools. Tools are organized by technical complexity, from browser-based to command-line interfaces which we’ve used in this course.

Browser-Based Tools

When to use: Quick prototyping, idea validation, no local setup required

Lovable

Full-stack app builder through conversation

  • What it builds: React frontend + Supabase backend
  • Speed: Hours to working MVP
  • Learning curve: Minimal
  • Link: lovable.dev

V0 by Vercel

UI component generator

  • What it builds: React + Tailwind components
  • Speed: Minutes per component
  • Learning curve: Minimal
  • Link: v0.dev

Bolt.new

Full-stack development environment in browser

  • What it builds: Complete applications with deployment
  • Speed: Hours to working prototype
  • Learning curve: Low
  • Link: bolt.new

Replit

Browser IDE with AI assistant

  • What it builds: Multi-language projects
  • Speed: Varies by project
  • Learning curve: Low to medium
  • Link: replit.com

Desktop IDEs

When to use: Professional development, complex projects, team collaboration

Cursor

AI-first code editor built on VS Code

  • Key features:
    • Composer: Multi-file editing
    • Tab: AI autocomplete
    • Chat: Conversational coding
  • Best for: Technical PMs and engineers wanting AI assistance
  • Learning curve: Medium
  • Link: cursor.sh

Windsurf

Agentic IDE with autonomous refactoring - Key features: - Cascade mode: Large-scale multi-file changes - Deep codebase understanding - Autonomous agent capabilities - Best for: Developers comfortable with AI making bigger changes - Learning curve: Medium to high - Link: codeium.com/windsurf

GitHub Copilot

AI pair programmer that works in any editor

  • Key features:
    • Inline code suggestions
    • Works in VS Code, JetBrains, Vim, etc.
    • Context-aware completions
  • Best for: Anyone who codes regularly
  • Learning curve: Low
  • Link: github.com/features/copilot

Command Line Tools

When to use: Maximum control, automation, transparent workflows

Claude Code (CLI)

Terminal-based agentic development tool - Key features: - Plan mode: Understand before changing - MCP servers: Extend capabilities - Project memory: Persistent context - Best for: Technical PMs, terminal-comfortable developers - Learning curve: Medium to high - Install: npm install -g @anthropic-ai/claude-code

Essential commands:

Command Action
Shift + Tab Toggle plan mode
/clear Clear context
# Add to project memory
@filename Reference files

Aider

Git-aware AI coding assistant

  • Key features:
    • Automatic git commits
    • Multiple AI model support
    • Git workflow integration
  • Best for: Developers who live in terminal
  • Learning curve: Medium
  • Link: aider.chat

Cline

VS Code extension with agentic capabilities - Key features: - Claude reasoning in VS Code - Web browsing - File editing and command execution - Best for: VS Code users wanting Claude - Learning curve: Low to medium - Install: VS Code extension marketplace

GitHub Copilot CLI

AI-powered command line assistant - Key features: - Natural language to shell commands - Git command suggestions - Explains what commands do - Best for: Learning terminal commands, writing complex bash scripts - Learning curve: Low - Install: gh extension install github/gh-copilot

Gemini CLI

Google’s AI assistant in terminal - Key features: - Multimodal AI (text, images, code) - Google Search integration - Long context window - Best for: Research-heavy tasks, using Google’s AI models - Learning curve: Low to medium - Install: npm install -g @google/generative-ai-cli

GitHub Codex

AI code generation and completion - Key features: - Natural language to code translation - Multiple programming languages - Code explanation and refactoring - Best for: Quick code generation, learning new languages - Learning curve: Low - Note: Powers GitHub Copilot

Supporting Services

Backend & Database

Supabase - Open source Firebase alternative - PostgreSQL database - Instant REST and GraphQL APIs - Authentication and storage - Real-time subscriptions - Link: supabase.com

Frontend Tools

Stitch by Google - Prompt-to-UI component generator - Generate React components from natural language - Link: stitch.withgoogle.com

Presentations

Gamma - AI presentation builder - Generate slide decks from prompts - Good for pitch decks and updates - Link: gamma.app

AI Chat Interfaces

For research, brainstorming, and code assistance:

Tool Best For
ChatGPT General tasks, coding help, broad knowledge
Claude Complex reasoning, long documents, coding
Perplexity Research with citations, current information
Gemini Google workspace integration
DeepSeek Open source alternative

Programming Languages

Languages with strong AI tool support:

Language Primary Use AI Support
JavaScript/TypeScript Web development (frontend and backend) Excellent
Python Backend, data processing, AI/ML Excellent
HTML/CSS Web structure and styling Excellent
SQL Database queries Very Good
Note

With AI coding assistants, you focus on what you want to build, not memorizing syntax. Describe your intent in natural language and let AI generate the code.

MCP Servers

Model Context Protocol extends Claude Code capabilities. Install as needed:

Playwright - Browser Automation

claude mcp add playwright npx @playwright/mcp@latest

Enables: - Automated browser testing - Screenshot capture - Web interface interaction

Common MCP Servers

  • File system - Access files outside project
  • Database - Direct database connections
  • API - External service integration
Tip

Start simple. Only add MCP servers when you need specific capabilities. Each server adds complexity.