Vibe Coding & AI Builders
Building sites by describing them: prompts, coding agents, AI builders and how to review what they write.
Agentic workflow
An agentic workflow is a process where AI agents carry out multi-step work, planning, acting, checking and handing off, with people approving key decisions.
AI agent
An AI agent is an AI system that pursues a goal over several steps, deciding what to do, using tools like a browser, terminal or APIs, and checking its own results.
AI website builder
An AI website builder generates a website from a description: it writes the copy, picks a layout and design and often publishes the result for you.
Bolt
Bolt is an in-browser AI development environment that generates, runs and deploys full-stack web apps from prompts.
Claude Code
Claude Code is Anthropic's agentic coding tool that works in a terminal, IDE or desktop app, reading a codebase, editing files and running commands to complete development tasks.
Coding agent
A coding agent is an AI that works in a real codebase: it reads files, writes and edits code, runs builds and tests, and fixes what fails.
Context window
The context window is how much text an AI model can consider at once, including your instructions, the files it reads and the conversation so far.
Cursor
Cursor is an AI-first code editor built on VS Code that combines autocomplete, chat and an agent that can edit multiple files in a project.
GitHub Copilot
GitHub Copilot is an AI assistant built into code editors and GitHub that suggests code, answers questions and can carry out multi-step coding tasks.
Hallucination
A hallucination is when an AI states something false with confidence, such as a function that does not exist, a wrong fact or an invented source.
LLM
A large language model is an AI system trained on vast amounts of text to predict and generate language, including code, which powers chatbots and coding agents.
Lovable
Lovable is an AI app builder that turns chat prompts into full-stack web apps with a React front-end and a hosted back-end, aimed at non-developers.
MCP
MCP, the Model Context Protocol, is an open standard for connecting AI assistants to external tools and data, such as databases, design files, browsers and APIs.
One-shot generation
One-shot generation is getting a complete, working result from an AI in a single request, without follow-up corrections.
Prompt
A prompt is the instruction you give an AI model: the request, the context and any constraints that shape what it produces.
Prompt engineering
Prompt engineering is the practice of writing and structuring instructions to AI models so they produce accurate, useful and consistent results.
Prompt to website
Prompt to website is generating a complete site from a written description, where the description decides the structure, copy and style of what gets built.
Replit Agent
Replit Agent is an AI agent in the Replit online IDE that builds, runs and deploys apps from a plain-language description.
Reviewing AI-generated code
Reviewing AI-generated code means checking what an AI wrote before it ships: that it does what was asked, handles errors, stays secure and does not quietly break other things.
System prompt
A system prompt is the standing instruction that sets an AI's role, rules and style for a whole conversation, before any user message.
Tokens
Tokens are the chunks of text, often word fragments, that AI models read and write; usage, limits and pricing for AI services are measured in tokens.
v0
v0 is Vercel's AI tool that generates React and Next.js interfaces and apps from prompts, designed to hand off into a real codebase.
Vibe coding
Vibe coding is building software by describing what you want to an AI in plain language and iterating on what it produces, rather than writing the code line by line yourself.