vibe codingvibe codingvibe coding

NAME

Vibe coding: building software by describing what you want in plain language and letting AI write the code. You guide the direction; the AI handles the syntax.

VibeCoding.app is the full-stack directory for the ecosystem: tools, comparisons, alternatives, skills, plugins, playbooks, workflows, agencies, events, in-depth reviews and editorials, best-of guides, and vibe debugging (the problems vibe-coded apps ship with, and the fixes).

SYNOPSIS

$ vibe code [intent] [stack?] [constraints?]
  → AI writes the code
  → You steer the direction
  → Ship before the weekend's out

Write a clear prompt, accept the diff, intervene when the model drifts, iterate. Solo founder on a first MVP or senior developer integrating AI into a production codebase: the workflow is the same, only the tool you reach for changes.

ORIGIN

The term was coined by Andrej Karpathy in February 2025. It stuck because it named something developers were already doing: pair-programming with a large language model, accepting the diffs, intervening when it drifts. Within months it had a name, dedicated subreddits, conferences, weekly tool launches, and a community of builders shipping production apps in plain English. The vocabulary stabilised around three categories of tools, which became this directory’s structure. (Origin card on the left has the tweet.)

THREE FLAVOURS OF VIBE CODING

WHY THIS DIRECTORY EXISTS

The landscape changes weekly. New tools launch, existing ones pivot, pricing shifts overnight. Developers waste hours evaluating tools that don’t fit their workflow, or worse, commit to a platform only to hit its limits months later. VibeCoding.app is the biggest and most comprehensive directory for vibe coding on the web, built to make that choice faster and better-informed.

How we work: every entry is researched and written deliberately. Structured Grok and SERP research packs feed each draft (sources verified, claims cross-checked, copy humanized) before it ships. We publish weekly and re-check tools when their state changes, pricing pivots, model drops, feature launches. A healthy queue of submissions keeps the pipeline full.

We don’t pretend to retest every tool every week. That’s an agent-scale problem, and the agents that can do it well aren’t shipped yet. Until then, depth where it matters beats breadth-for-its-own-sake. We’re independent: editorial judgement stays separate from commercial relationships, and featured placements are always disclosed.