Vibe coding is the practice of building software by telling an AI tool what you want in plain language, then shaping the result by feel instead of writing every line by hand. The term was coined by Andrej Karpathy in early 2025, who described it as giving in to the vibes and letting the model handle the code while you focus on the outcome.
In practice, a vibe coder prompts a large language model, looks at what it produces, runs it, and asks for changes in a back-and-forth loop. This lowers the barrier to building software, so non-engineers can ship working prototypes and experienced developers can move a lot faster. The tradeoff is that code you do not fully understand can be harder to debug, secure, and scale, which is why human review still matters.
Examples of vibe coding tools:
Vibe Coding Explained:
In this video, Saj Adib from Skills Leap AI walks you through how to build a fully functional YouTube content system using 'vibe coding' - all without writing a single line of code. You'll see the exact step-by-step process of using plain language to prompt, refine, and build a custom app from scratch.
FAQs
Vibe coding is a way of building software where you describe what you want to an AI tool in everyday language and let it generate most of the code. Instead of writing each line yourself, you guide the AI through prompts, testing, and iteration until the result works.
AI researcher Andrej Karpathy, a co-founder of OpenAI and former AI lead at Tesla, coined the term in early 2025 in a widely shared post. The phrase spread quickly from social media into mainstream tech coverage.
Vibe coding is great for prototypes, internal tools, and getting an idea off the ground fast. For production systems, teams still need to review the generated code for logic, security, and scalability, since shipping code nobody fully understands can create problems later.
