A new Claude Code plugin is making waves on Hacker News for tackling one of the most common gripes with AI coding assistants: their tendency to respond in overly formal, verbose "AI-speak." The tool, called claudish-to-english and developed by GitHub user gvzdv, automatically rewrites each assistant message into more conversational English that developers can actually parse at a glance.

What the Plugin Does

The core functionality is straightforward but clever. Rather than letting Claude's responses slide by in their typical dense, hedged-with-caveats style, claudish-to-english intercepts each message and outputs a parallel "plain English" version alongside it. Think of it as having a translator sit next to you while pair programming with an AI that never learned how to talk to humans.

Why This Matters for Developer Workflows

Claude Code, Anthropic's CLI tool for AI-assisted coding, has gained serious traction among developers who want to automate repetitive tasks or get second opinions on their code. But the assistant's default communication style can be exhaustingβ€”every response wrapped in disclaimers, every suggestion prefixed with "I should note that..." For developers working through a complex debugging session or trying to maintain flow state, wading through verbose AI output is a distraction nobody asked for.

The Hacker Community Responds

The post picked up modest engagement on Hacker News, with readers debating whether this kind of translation layer represents a band-aid fix or genuine UX improvement. Some commenters argued that developers should push back directly on how Claude Code formats its responses rather than layering another tool on top. Others saw value in keeping the original verbose output for auditability while using the plain English version for day-to-day comprehension.

Key Takeaways

  • The plugin targets a real pain point: AI assistants communicating in ways that interrupt developer flow
  • Available now on GitHub at gvzdv/claudish-to-english with installation via Claude Code's plugin system
  • Community reaction suggests this could be the first of many "AI output translation" tools as these assistants proliferate

The Bottom Line

We love seeing grassroots tooling that improves the human side of AI interactions. Whether this specific approach catches on or something more integrated becomes standard, the underlying problemβ€”AI outputs that respect model capabilities over human attention spansβ€”isn't going away. Kudos to gvzdv for shipping something useful instead of just complaining about it.