# Vibe coding

A practitioner label for software creation driven primarily by natural-language iteration and observed behavior, often with limited attention to the generated code.

technical-term · Operating model · practitioner · Reviewed 2026-08-09

## Definition

Vibe coding describes a highly exploratory mode in which a person asks a model to create or change software, runs it, and iterates from visible results. It can accelerate prototypes and learning, but consequential software still requires explicit intent, architecture, tests, provenance, security, ownership, and release evidence.

The phrase is culturally influential but informal, inconsistently defined, and not an engineering standard.

## Why it matters

The mode is valuable when reversibility is high and the objective is discovery.

Agentic engineering begins where organizations make the surrounding intent, context, execution, verification, and accountability system explicit.

## System anatomy

- **Natural-language steering:** The person directs changes through conversation.
- **Rapid feedback:** Visible execution drives the next request.
- **Low code inspection:** The person may rely more on behavior than detailed implementation review.



## Important distinctions

- **Agentic engineering:** Agentic engineering treats the full delivery and control system as the engineering artifact.
- **Prototyping:** Prototyping is a lifecycle intent; vibe coding is one informal interaction style.



## Implementation signals

- Keep use proportional to reversibility and impact
- Convert discoveries into explicit requirements and tests before production
- Do not hide generated-code provenance

## Failure modes

- Prototype assumptions reach production unexamined
- The operator cannot explain, maintain, or verify the resulting system



## Related knowledge

- [Agentic engineering](https://aisdlc.ai/agentic-engineering/agentic-engineering) — The discipline of designing intent, context, memory, tools, execution, verification, control, and evidence so goal-directed agents can perform meaningful work while humans remain accountable.
- [Agentic software factory](https://aisdlc.ai/agentic-engineering/software-factory) — An operating model that assembles reusable agents, workflows, environments, specifications, checks, and delivery controls into a repeatable software-production system.
- [Intent engineering](https://aisdlc.ai/agentic-engineering/intent-engineering) — The practice of turning goals into versioned specifications, constraints, interfaces, invariants, decision rights, acceptance criteria, and testable outcomes before agents implement them.
- [Evidence engineering](https://aisdlc.ai/agentic-engineering/evidence-engineering) — The design of versioned, queryable evidence linking requirements, decisions, implementations, tests, evaluations, approvals, deployments, runtime signals, and lifecycle actions.

## Sources and further study

- [Andrej Karpathy — Software Is Changing (Again)](https://www.youtube.com/watch?v=LCEmiRjPEtQ) — A thought-leadership lens on natural language, neural networks, and changing software interfaces—not a formal standard. **Use:** Practitioner perspective. **Limitation:** This is an attributed thought-leadership or practitioner source, not a consensus standard or independently validated general law.
- [GitHub — Spec-driven development with AI](https://github.blog/ai-and-ml/generative-ai/spec-driven-development-with-ai-get-started-with-a-new-open-source-toolkit/) — A structured specification-to-plan-to-task workflow for making intent durable and actionable by coding agents. **Use:** First-party case study. **Limitation:** This first-party account documents one organization, product, or implementation context and should not be generalized without local evidence.
- [OpenAI — Harness engineering: leveraging Codex in an agent-first world](https://openai.com/index/harness-engineering/) — A production case study in designing repositories, tools, feedback loops, and environments for agent-first delivery. **Use:** First-party case study. **Limitation:** This first-party account documents one organization, product, or implementation context and should not be generalized without local evidence.

---

This library synthesizes cited research, standards, official documentation, and clearly attributed practitioner perspectives. Maturity describes the state of a concept—not vendor endorsement, production readiness, or permission to deploy. Benchmarks and demonstrations do not replace use-case evaluation, governed controls, independent verification, or named human release authority.