// AI WRITING UTILITY ALTERNATIVE

The Claudette alternative for AI-tell in published content, not just one terminal reply

Claudette is a free skill that de-buzzes one Claude reply in your terminal. Kompozy governs AI-tell across 18 formats and publishes to 9 platforms.

Last verified · 2026-08-21 · by Moe Ameen

If you searched "Claudette alternative," start with what Claudette actually is, because it decides whether you are in the right place. Claudette — really the /debuzz skill in Adnan Akil's open-source "nobuzz" project — is a free Claude Code command that takes Claude's verbose, clickbait-style reply and reruns it through the Google Gemini CLI to rewrite it in plain English. It is a clever, tightly-scoped developer utility for one job: making a single message in your terminal readable. If that is all you want, Claudette is excellent and nothing here replaces it.

But a lot of people reach for "an alternative" because they hit the tool's edges. Claudette fixes AI-tell one reply at a time, after the fact, and only in your terminal. It has no memory of your brand voice, produces text only, and cannot publish anything. I run Kompozy, so weigh that, but the honest framing is that these are different categories: Claudette is a point fix for Claude's chat output; Kompozy is a generation-and-publishing engine that keeps clickbait out of the content you actually ship.

So the real question is your bottleneck. If it is "make this one Claude answer sound normal," Claudette is the right, free tool. If it is "generate a week of on-brand video, images, carousels, a blog, and a newsletter with no AI-tell, and get them onto every platform on a schedule," a terminal skill is the wrong shape, and this page is about the tool that is the right one.

Everything below is reconciled against the nobuzz README and Kompozy pricing from our own page, checked on 2026-08-21. Where Claudette is the better tool for a job, this page says so.

What Claudette (nobuzz / the /debuzz skill) does

Claudette installs into Claude Code as the /debuzz slash command. When Claude answers in its theatrical register — framing a simple bug as a "load-bearing assumption" with numbered revelations — you run /debuzz and it captures that reply, pipes it through the Gemini CLI with plain-English instructions, and prints Gemini's rewrite. Using a second model rather than asking Claude to fix its own prose is the deliberate design: the project's premise is that prompting Claude alone rarely removes the style for long. It ships three modes — colleague (default, keeps all code and paths, zero theatrics), manager (about a third the length, no code), and director (three to five executive sentences) — and also triggers on phrases like "say that in normal english." It is MIT-licensed and free. What it does not do is anything past that one rewrite. There is no persistent brand-voice profile, no image or video output, no captions, no per-platform sizing, no review step, and no scheduling or publishing. It also is not self-contained: it needs the Google Gemini CLI installed and authenticated, so the de-buzzing depends on a second model and a second account. It is a terminal-quality-of-life tool, not a content workflow.

Why people look for a Claudette (nobuzz / the /debuzz skill) alternative

You look past Claudette the moment "fix one Claude reply" stops being your bottleneck and "ship on-brand content" becomes it. Claudette treats AI-tell as something to remove after generation — Claude writes it, Gemini rewrites it flatter. For a single terminal message that is fine. For a publishing cadence it means running every caption, script, and draft through a second tool by hand, with no guarantee the rewrite matches your specific voice, because Claudette has no brand profile to enforce. There is also a dependency cost people underweight. Claudette is free, but it only works if you install and authenticate the Gemini CLI, so you are maintaining a second model just to translate the first one's output. And it still produces plain text — you need separate tools for images, video, captions, scheduling, and publishing, plus the glue to connect them. The alternative worth considering is not another de-buzzing script; it is an engine where voice is governed at the moment of generation across every format, so there is no AI-tell to strip out, and where review and multi-platform publishing already come in the box. That is Kompozy.

Claudette (nobuzz / the /debuzz skill) vs Kompozy — feature comparison

FeatureClaudette (nobuzz / the /debuzz skill)KompozyNote
Removes AI-tell / clickbait styleYes — post-hoc, by rerunning text through GeminiYes — governed at generation by a Persona Brief and banned-word filterClaudette translates after the fact; Kompozy prevents the style at the source.
Works without a second model / APINo — requires the Gemini CLI, authenticatedYes — managed Claude and OpenAI are built inClaudette de-buzzes Claude by depending on Google Gemini.
Persistent brand-voice profileNo — no memory of your voice between rewritesA Persona Brief defines your register once and applies it to every asset
Rewrites arbitrary pasted textYes — /debuzz [mode] [text]Yes — ingests a source and regenerates it in your voice
Output formats beyond textNo — text only18 formats across video, image, carousel, blog, and newsletter
Branded captions / per-platform reframeNoBurns in captions and sizes 9:16 / 1:1 / 16:9 per destination
Pre-publish review gateN/A — no publishingPer-post review pipeline before Autopilot schedules
Multi-platform publishing + schedulingNoPublishes to 9 platforms plus Mailchimp and GHL/WordPress with autopilot
Consistent recurring persona / face-lockNoAn AI Influencer persona pool with Gemini face-lock keeps one identity across posts
PriceFree, open-source (MIT); you supply Gemini accessPaid credits, with a bring-your-own-key option on the Founding tier

Pricing — Claudette (nobuzz / the /debuzz skill) vs Kompozy

TierClaudette (nobuzz / the /debuzz skill) planClaudette (nobuzz / the /debuzz skill) priceKompozy planKompozy price
EntryClaudette (open-source skill)Free (MIT) — you supply a Gemini API key or CLI authStarter$99/mo (5,500 credits)
MidClaudette + Gemini API usageFree skill; you pay Google for Gemini API callsPro$299/mo (18,000 credits)
TopRoll-your-own style stackFree tools plus your time to wire generation and publishingEnterpriseCustom (sales-led)
Pricing verified 2026-08-21from each vendor’s public pricing page. Promotional rates rotate monthly — verify before purchase.

What Claudette (nobuzz / the /debuzz skill) does well

  • Free and open-source (MIT), and installs in a couple of commands into Claude Code.
  • Solves a real, specific annoyance — Claude's theatrical engineering prose — rather than a vague one.
  • Smart design: it uses a second model (Gemini) to translate, instead of asking Claude to de-buzz its own writing.
  • Three well-chosen modes — colleague, manager, director — map cleanly to real audiences.
  • Low friction: triggers on natural phrases like "say that in normal english," not just a command.
  • Rewrites any pasted text, not only Claude's previous reply.

Where Claudette (nobuzz / the /debuzz skill) falls short

  • Requires the Google Gemini CLI installed and authenticated — a second dependency and a potential second bill.
  • Post-hoc by design: it fixes output after it is written instead of preventing the style at the source.
  • Scope is your terminal — it de-buzzes one message, not a content pipeline.
  • Text only; no images, video, captions, scheduling, or publishing.
  • No persistent brand-voice profile, so it cannot enforce your specific voice across a batch of content.
  • The author frames it openly as a workaround, not a durable fix for how Claude writes.

Pick Claudette (nobuzz / the /debuzz skill) when…

  • You just want your terminal output readable. For making one Claude Code answer plain and skimmable, Claudette is free, instant, and exactly the job.
  • You work primarily inside Claude Code. It lives right where the theatrical prose appears, so there is no context-switch to a separate app.
  • You want a free, hackable tool. It is MIT-licensed and only a few files, so it is easy to fork and adjust to your taste.
  • You already run the Gemini CLI. If Gemini is already set up, Claudette adds no new dependency and costs you nothing.

Pick Kompozy when…

  • You need AI-tell gone from published content. A Persona Brief and banned-word filter govern every generation, so clickbait style never reaches a caption, script, or blog.
  • You want it prevented, not translated. Kompozy enforces voice when the text is written, so there is no second model to run over the output afterward.
  • You publish across platforms. Kompozy schedules and fans content to 9 platforms plus email and blog with autopilot; Claudette has no publishing at all.
  • You need more than text. Kompozy generates 18 formats of video, image, and long-form; a de-buzzing script produces only rewritten text.

Why Kompozy is the Claudette (nobuzz / the /debuzz skill) alternative we recommend

Kompozy is a full AI content generation and 9-platform publishing engine, not a de-buzzing script. Claudette is a genuinely clever one — free, open, and well-designed for its narrow job — but that job is a single reply in your terminal, fixed after the fact by a second model. The gap between "this one Claude answer reads normally now" and "I have on-brand posts live across every platform with no AI-tell" is the entire content workflow, and that gap is what Kompozy fills.

From one input, Kompozy generates 18 output formats — HeyGen avatar Persona Shorts, fal.ai VFX hooks, face-locked Persona Photos, brand-exact carousels, quote cards, blog articles, and email newsletters — all in a voice defined once by a Persona Brief with a banned-word filter, so there is nothing to translate later, all routed through a per-post review gate, then scheduled and published to Instagram, Facebook, TikTok, YouTube, LinkedIn, X, Pinterest, and Threads, plus Mailchimp and GHL/WordPress, on autopilot. Keep Claudette for your Claude Code sessions; let Kompozy handle the content your audience actually sees. And where Claudette leans on Gemini to fix Claude's prose, Kompozy runs managed Claude and OpenAI directly, with a bring-your-own-key option on the Founding tier — one system, no second model to babysit.

Frequently asked questions

Is Kompozy a replacement for Claudette?

Not a like-for-like one — they are different categories. Claudette is a free Claude Code skill that de-buzzes one reply in your terminal by rerunning it through Gemini; Kompozy is a generation-and-publishing engine that keeps AI-tell out of the content you publish and ships it across 9 platforms. If you only want a terminal reply cleaned up, Claudette is the closer fit. If you were trying to keep clickbait out of real content, Kompozy is the finished tool.

Can I use Claudette and Kompozy together?

Yes, and it is a clean split. Use Claudette in Claude Code to keep your working notes and explanations plain, then drop any idea worth publishing into Kompozy as a source. Kompozy regenerates it in your Persona Brief voice with no AI-tell, produces a full multi-format batch, runs each asset through a review gate, and publishes on a schedule. Claudette fixes your terminal; Kompozy makes and distributes your content.

Does Claudette cost anything?

The skill itself is free and MIT-licensed. The catch is that it de-buzzes Claude by piping the text through the Google Gemini CLI, which you must install and authenticate — so you depend on a second model and, at volume, a second bill. Kompozy governs voice at generation with managed Claude and OpenAI, so there is no separate translation step to pay for.

Is it actually called Claudette?

Not officially. The real project is "nobuzz" and the command is /debuzz; the README jokes that it is "definitely absolutely not called 'Claudette'." People use the nickname anyway, including in the Hacker News thread where it circulated in August 2026.

How does Kompozy stop AI-tell if Claudette needs a second model to do it?

Claudette needs Gemini because it is translating text Claude already wrote in its clickbait register. Kompozy never writes in that register in the first place: a Persona Brief and banned-word filter constrain the generation, so the voice is correct at the source and no second model is required to clean it up. That is the difference between prevention and post-hoc translation.

Related deep guides

See Kompozy pricing · Get Started →