Video Review
Unvibe: Explain AI-Written Code in Plain English, Free
Developers using AI-generated code can struggle to understand the output and make it their own. A Mac desktop learning layer helps bridge this gap by explaining AI-generated code in the context of a developer's repository and workflow. It allows users to select code, choose an explanation depth, and learn what it does, why it works, and how it connects to their project without leaving their workflow. The layer is designed to remain available across a developer's workflow, rather than becoming another separate place to work. The application offers five explanation depths, letting users control the level of detail they receive, from purpose and syntax to execution flow, architecture, risks, and tradeoffs. It also provides repository context, moving outward from the selected expression to the surrounding function, file, dependencies, Git change, project structure, and underlying concept. Users can save explanations, track their history, and participate in comprehension checks. The tool is intended for developers who use AI-generated code and need to close the ownership gap before shipping, debugging, securing, and maintaining it. It is currently in a private Mac beta with invite-only access and no credit card required. The free version allows users to understand selected code with 50 explanations each month, core explanation depths, saved explanations, and study. A Pro version adds broader project context, including Git diffs. The application handles explanation requests and filters likely secrets locally before sending approved context to a model, prioritizing user privacy.