Lacquer

Lacquer rewrites AI text to human standards.

For a while, AI text was enough. You fed a model your thoughts, points, and conclusions, and it whipped up X posts, LinkedIn updates, and thought pieces to your exact specifications. All was well.

Now people can tell. The public is rapidly inoculating itself to the characteristic tone and style of various LLM models. Anyone who can't has access to detection tools and watermarking that work better every day.

No matter how good your ideas are, using AI to express them is turning off a significant percentage of your audience.

To fix that, you need a human airgap: an expert writer who can rework your AI-generated piece into a human voice that doesn't vibe as AI, because it isn't AI.

Send us text Read the examples first

How it works.

  1. First, you work with your preferred LLM tooling and your own deep expertise to make the article only you can, exactly the way you want it. This eliminates courier problems — we know precisely what you’re trying to say.
  2. After that, you send the completed piece to us. We do a full rewrite, replacing all the prose while leaving your structure, ideas, thoughts, and conclusions untouched.
  3. You save the writing time and get the best possible outcome — your expertise expressed through expert, high-end human writing.

We refine text.

Our writers don’t edit. They recreate, preserving the structure, flow, and message intended while removing AI tells and enforcing human taste and style.

Generated

A friend of mine ran reactors on a fast-attack submarine. I asked him once what the actual discipline was — the thing underneath all the procedure — and he said something I’ve thought about every week since.

He said the point of reporting a two-degree anomaly isn’t the two degrees. The point is that a crew that reports two degrees at 3 a.m. on a Sunday is a crew that still has the reflex. You don’t maintain the reflex for the thing that matters. You maintain it on the things that don’t, so it’s there when something does.

We had let the reflex atrophy. Ninety seconds of atrophy, nine days deep.

Rewritten

I have a friend who once ran the reactors on a fast-attack submarine. I asked him what it was like to work in that environment, one where a small lapse in judgement could mean disaster.

He said it was all about reporting two-degree anomalies. Those, he explained, are next to nothing in the abstract. But a crew in the habit of reporting two-degree problems is proving it will catch and report real problems, too.

Surviving, he said, was a matter of building reflex via discipline. The exact same kind of discipline, I think, that both my team and I had let slip.

One of these examples pings 100% AI on Pangram. One doesn’t. You can guess which is which.

Why not Fiverr?

If you've tried that option, you already know why it doesn't work. If you haven't, go try now; we'll wait.

Back? Good. What you just experienced was some combination of four problems:

Scams

Most writers on Fiverr are leveraging AI. They will take your money and reprocess your article through the same models you use, leaving you with the same set of problems.

They can't do human rewrites because they aren't human writers in that sense. They don't have the capability to begin with.

Understanding

To rework the prose around a set of structured ideas without lossy translation problems, a writer has to understand what you are trying to say.

Our team includes a respected novelist, a noted blogger, and a corporate content specialist who has worked as a creative inside Series A and B startups. We get what you are saying in a way Fiverr talent can't.

Quality

Your bar is high. At minimum, a rewrite has to meet AI quality standards. Ideally it exceeds them. We can do this; most can't.

Turnaround time

Some conversations won't wait. Our team turns around most requests in hours, not days, and without the need to sift through contractors looking for availability that fits your timeframe.

Pricing.

50¢ per word

Standard queue. Most work comes back the same day.

$1 per word, rush

Straight to the front of the queue.

50-word minimum on all jobs.

Send us text.

0 words / —

Over 8,000 words? Upload a file instead.

.txt, .md, .docx or .pdf. 10 MB max. We work from extracted text, so your original file is never opened on anyone’s desktop.

You’ll see a confirmation page with the final word count and price before anything is charged.

What happens to your text. It's stored encrypted, visible only to the writer working on it, and deleted 30 days after the job closes. It is never used to train anything and never shared outside the team. If you want it gone sooner, ask and it's gone that day.

Questions.

Can you always beat Pangram and other detectors?

Almost. Because we rewrite every sentence and paragraph, we are in very real terms presenting detectors with 100% human text.

Good, high quality detectors (like Pangram) are very good at avoiding false positives. Other lesser products often flag human writing as AI in error; we can’t prevent that.

As a precaution against the errors we can avoid, we purchase and include a Pangram report showing 0% AI involvement with every finished piece of sufficient length — usually 50 words or more.

Why do I need to submit AI-drafted source material?

We can never be more expert than you as it relates to your field of expertise.

Conventional ghostwriting tries to bridge that gap by having you provide a rough version of what you’d like to say, but it’s lossy. Points are often misunderstood or missed entirely. Nuance is lost. More often than not, the finished work feels non-expert in a way that renders it useless.

By writing your piece with AI in advance, you can ensure it includes everything you want before we rewrite it at a world-class, human level. It ensures the best mix of your knowledge and our skill.

Can you do more than rephrase? Could you help me refine my thoughts, or write an entire article from scratch?

We can, but it involves a longer conversation (and, to be upfront, a much higher price). Email us and we’ll talk it over.

What about confidentiality and copyright?

We explicitly disclaim all rights to your work except the momentary permission you give us to work on your piece. In all other aspects, both the submitted source material and the finished work remain yours to use how you wish. We use it for nothing, and delete both records after thirty days.

Unless you give us explicit permission that indicates otherwise, we don’t disclose what we write or who we write for under any compulsion weaker than a subpoena.

What kind of work do you take? Is there anything you won’t do?

We can humanize almost anything, including social media posts, blog posts and thoughtpieces, emails, and works of fiction.

We are politically viewpoint neutral in most ways that matter.

Highly technical work, like hard-science academic papers, can sometimes be beyond us. Email us with what you need and we’ll let you know what we can and can’t handle.

We do not work with erotica and pornographic material.

We do not help students cheat on their homework.

If you think you have a good reason for us to make an exception, email us.

Can you be engaged via a contract or subscription, rather than piece-by-piece at per-word pricing?

It’s possible, but unlikely. Email us and we’ll have a conversation about it.