Skip to main content
Lornets

Evidence note

Vibe Coding in Software Development: A Multivocal Literature Review

Preprint / Working Paper2026Preprint

What was examined?

A multivocal review of 47 sources, comprising 28 peer-reviewed sources and 19 grey-literature sources, covering research and practitioner evidence on vibe coding.

Multivocal literature review covering literature from 2022 through October 2025.

47 included sources.

Key findings

  1. 01Short-term productivity or time-to-prototype gains were reported in 21 of the 47 included sources.
  2. 02Evidence about production use and long-term maintainability was substantially thinner than evidence about prototyping and early-stage development.
  3. 03Among sources discussing safeguards, validation and testing pipelines were the most frequently identified safeguard category.

Why it matters. Lornets interpretation.

Rapid software creation and demonstrated production suitability are different questions. The evidence base is currently much stronger around creation and prototyping than around long-running production operation.

This is the Lornets reading of the source, not a finding of the source itself.

What it does not establish

  1. 01The evidence base includes grey literature as well as peer-reviewed research.
  2. 02Explicit production evidence was limited.
  3. 03The review does not establish that vibe-coded software is inherently unsafe.
  4. 04It does not establish that migration or rewriting is generally necessary.

Source

Venue
Preprint
Evidence type
Preprint / Working Paper
Published
2026-07-22
Status
Current

Read the original research

Relevant Lornets framework areas

Framework domains

  • Architecture & Maintainability
  • Delivery & Change Control

Related evidence

Source record

Published
2026-07-22
Last verified
2026-08-11
Source status
Current
Back to Research & EvidenceRead the Production Assurance Methodology