Reference
A reference for vibe coding
Vibe coding is the practice of building software by describing requirements in natural language and letting an AI generate and iterate the code.
The developer steers this process by defining intent and reviewing results rather than writing every line of code manually.
This reference covers the definition of vibe coding, the tool landscape, and the trade-offs regarding risk, cost, and ownership. All claims on this site are tied to sources you can check.
Contents
The six parts of this reference
01
What vibe coding is
Definitions of vibe coding and its origins in software development.
02
How it works
The mechanics of prompting, context, model interaction and code review loops.
03
The tool landscape
A survey of the current tools and platforms used for vibe coding.
04
Best practices
Working practices that help achieve reliable results when building with AI.
05
Risks and limits
The failure modes, security vulnerabilities and maintenance challenges of this approach.
06
For businesses
How companies can adopt vibe coding while managing team and project requirements.
Unfamiliar with a term? Read the glossary. Definitions of the technical vocabulary used throughout this site.
Latest writing
Recent notes
How to review AI-generated pull requests as a team
AI writes faster than your team can read. How to structure a review for AI-generated pull requests: disclosure, stated intent, size caps, a fixed reading order, and triage by blast radius.
AI code security and the versions a model invents
AI assistants invent version numbers that were never published, and a quarter of upgrade recommendations in one 2026 study referenced versions that do not resolve. How to catch them before they reach a lockfile.
How AI coding tools handle context window limits
Every AI coding tool is really a context manager with a model attached. How selection, indexing and compaction work inside Cursor, Claude Code, Copilot and Codex, and why a million-token window did not end the problem.