GlossaryModelsNew to AI

Vibe coding

Also written: Vibe-code, Vibecoding

Coding by describing what you want in natural language, with AI generating it.

In plain terms

Vibe coding is a way to create software by telling an AI what you want it to do, instead of writing traditional code. You use plain English, like describing a feeling or a goal, and the AI tries to build the program based on your description.

Think of it like asking a friend to build something for you. You don't need to know the exact steps or tools; you just explain the final result you're aiming for. For example, you might tell an AI, 'Make a simple game where a cat jumps over obstacles,' and the AI would generate the game code for you.

This approach is useful because it lowers the barrier to creating software. People who aren't professional programmers can start building applications by simply describing their ideas.

The expert version

Vibe coding refers to the paradigm of generating software through natural language prompts, where the AI interprets the user's intent, or 'vibe,' to produce functional code. This approach leverages large language models (LLMs) to translate high-level descriptions and desired outcomes directly into executable programs, reducing the need for explicit, low-level programming syntax.

Tradeoffs include potential for less precise control and the need for iterative refinement to align the AI's output with complex requirements. However, it significantly accelerates prototyping and democratizes software creation for a wider audience.

Why it matters now

Vibe coding is appearing now as AI models become more capable of generating functional code from natural language prompts. Recent developments show AI assisting in building games and complex applications, suggesting a shift towards more intuitive and less structured coding methods, especially with the rise of agentic engineering.

Where it shows up 17 items, newest first
How each company uses it
Anthropic Claude
Anthropic's recent work suggests a focus on agentic planning and rebuilding skills for their models, aligning with more intuitive approaches to AI-assisted development.
Google Gemini
Google's recent mention of a 'vibe coding course' with Kaggle indicates a large-scale effort to train developers in building AI agents using natural language, simplifying the coding process.
OpenAI ChatGPT
OpenAI's advancements in agent swarms and tools like Codex Sites suggest a move towards making AI-generated code more usable and accessible, potentially enabling less structured development workflows.
Related terms

Prompt engineeringAI agentsCode generationNatural language processingLow-code developmentGenerative AI

Often confused with
Prompt engineering
Prompt engineering focuses on crafting effective inputs for
Low-code development
Low-code platforms use visual interfaces and pre-built com
Code generation
Code generation is the act of producing code, which vibe coding
Agentic programming
Agentic programming involves AI agents performing tasks aut
Origin

The term 'vibe coding' emerged from the AI community, particularly around the use of large language models for software development. It reflects the informal and intuitive way users can interact with AI to generate code based on desired outcomes or 'vibes'.

Try this about 5 minutes

Open a chat interface with a capable AI model like ChatGPT, Gemini, or Claude. Describe a simple application you want, for example, 'Write Python code for a basic calculator that can add, subtract, multiply, and divide.' Then, ask the AI to provide the code. You can then copy and run this code in a suitable environment.

Try it: Playbook entries that use thisPlaybook →

The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.

Questions people ask

Do I need to know how to code to do vibe coding?
Not necessarily. The goal of vibe coding is to allow people to create software by describing what they want in natural language, even if they don't know traditional programming languages.
Is vibe coding reliable for complex projects?
For simple to moderately complex tasks, vibe coding can be effective. However, for highly intricate or mission-critical software, traditional coding methods with rigorous engineering practices are often still preferred.

First seen on Subvolts Nov 18, 2025. Updated Oct 14, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.