ScriptGrain vs HyperWrite

ScriptGrain keeps a persistent, inspectable profile: 45 named attributes measured across everything you have published. HyperWrite matches whatever examples you paste, then forgets them.

The actual difference

HyperWrite matches the style of the examples you paste. ScriptGrain measures 45 named attributes across everything you have published and keeps that profile as an artefact you can inspect, refine with more samples, and re-check: the measurement is the product, not a step inside it.

HyperWrite's entry plan is $19.99/mo for three personas. ScriptGrain's Operator plan is £29/mo for three profiles, and Writer is £12/mo for one.

ScriptGrain is not a replacement for the assistant you already use. It measures your voice from writing you have published and hands that profile to ChatGPT, Claude, Muse or any MCP client through its connector, then scores every draft against it. ScriptGrain makes your AI better at being you.

How the connector works

Every AI voice tool compared: prices, scores and API access, checked

	ScriptGrain	HyperWrite
Entry price	£12/mo (Writer, 1 voice profile)	$19.99/mo
Free tier	Yes, See your own voice profile and try one draft. No card.	Referenced on the pricing page (limited monthly credits) but the tier's contents were not stated
Voices on entry plan	1 (3 on Operator, £29/mo)	3 Custom Personas (Premium)
How the voice is built	45 measured style attributes, extracted from writing you have already published	Custom Personas built from examples you provide; its style tool asks for two examples and matches their style and format.

HyperWrite figures read from their pricing page on 3 August 2026.

What HyperWrite is built for

HyperWrite is a writing toolkit that happens to have a voice feature. That ordering matters. Its positioning line is "Make the AI write like you, not like a robot", and Custom Personas are the core of that promise. But around the personas sits a much larger set of things: hundreds of one-click writing tools, cited research, and a browser extension that offers suggestions wherever you type.

So the product is built for breadth. You open it, pick a tool, get a draft, move on. If your day is a long run of small writing jobs (an email here, a rewrite there, a bit of research with sources), that shape suits you. The voice matching is one lever among many, and it's a lever you pull when you need it.

And to be fair, that's a sensible design. Most people writing at work don't need a measured model of their prose. They need a decent first draft that doesn't read like a press release from a robot. HyperWrite goes straight at that.

How HyperWrite handles voice

Here's how the voice side actually works, going by HyperWrite's own pages as read on 2026-09-20.

There are two single-use tools. "AI Write in My Style" takes exactly 2 writing examples plus a request, then drafts in that style. "AI Style Mimic" takes 1 reference text and 1 text to rewrite, and matches the rewrite to the reference. Both are example-driven. You paste, it imitates.

Then there are Custom Personas: saved persona profiles that get reused across the product. HyperWrite states no sample size for them, so you can't tell from the page how much writing a persona is meant to be built from. That's not a criticism, just a gap in what's published.

The entry plan, Premium, is $19.99 a month (or $16 annually, both read on 2026-09-20) and includes 250 AI messages and 3 Custom Personas. There's also a free plan with limited monthly credits.

What you won't find is a voice score. Personas are style templates, not measured against a profile. Nothing tells you how close a draft landed, or which features drifted. You judge by eye. For a lot of jobs, your eye is perfectly good. For a team of five people sharing one brand voice, it gets shaky fast. One person's "sounds right" is another's "a bit off", and there's no number to settle it.

We found no API or MCP server either, as of 2026-09-20. That limits where a persona can travel. It lives inside HyperWrite.

How ScriptGrain handles voice

ScriptGrain starts from a different object. Instead of examples you paste at the moment of need, it builds a profile from writing you have already published, and that profile is the thing you keep.

The profile holds 45 attributes in 8 layers: lexical, syntactic, tone and register, rhetorical, punctuation and format, function words, content patterns, and quirks and cadence. The attributes are concrete. Average sentence length and its variance. Contraction rate. Comma density. Semicolon and dash habits. Formality on a 0 to 10 scale. Confidence versus hedging. Opening style, argument structure (evidence-first, conclusion-first, narrative), paragraph openers, pronoun mix. The full list sits in the writing voice attributes reference.

The bar to start is low. One sample of 50+ words is enough to make a profile, though three or more pieces and about 3,000 words is the recommendation. Extraction usually takes under a minute; the median was 39 seconds when measured on 2026-09-28.

Every draft then gets scored from 0 to 1 against the profile. That's voice match. On-voice drafts typically land between 0.85 and 0.95. Partial matches sit from 0.5 to 0.8. A clearly different voice comes in under 0.45. The score arrives with per-feature deltas, so you see the draft's value next to the profile's value for each feature. The method is published in the voice measurement framework, which means you can check the working rather than trust it.

Two more pieces matter. Polish revises a draft toward a target score (default 0.9). And drafts are written with the profile as a constraint, with the app learning from the edits a user saves.

Then there's reach. The profile travels to ChatGPT, Claude, Gemini, Cursor, Meta Muse and any MCP client through the MCP server and the REST API, read live each time. Both are on every plan, Free included. It is never a pasted prompt. You can read more on the API and MCP server page.

The same job in each tool

Take one job and run it twice. A small consultancy needs a 600-word client update in the founder's voice. The founder writes in short sentences, uses contractions constantly, opens most paragraphs with a blunt claim, and never hedges.

In HyperWrite, you'd open AI Write in My Style (or a saved persona), paste two of the founder's past emails, describe the update, and generate. The draft comes back quickly. It will likely pick up the surface feel: the brevity, the casual register. You read it, tweak a few lines, done. Fast, cheap, and good enough for plenty of purposes.

The catch shows up on the second update, written a month later by a different team member. The two examples they paste are a different two. The persona, if saved, carries some continuity, but nothing states what it carries or how closely. Whether the new draft matches the first one is a judgement call.

In ScriptGrain, the founder's published writing is analysed once. The profile captures the sentence-length variance and the contraction rate, plus the conclusion-first structure. The team member drafts the update (in ChatGPT, say, reading the profile live through MCP) and the draft gets scored. Suppose it comes back in the partial-match band, 0.5 to 0.8. The deltas point at why: sentences running too even, contractions too sparse. Polish pulls it toward the 0.9 target, or the writer fixes those two things by hand and re-scores.

Same job. One route gets you a draft quickly and leaves the verdict to you. The other gets you a draft plus a reason it is or isn't on voice. Which of those you want depends on how many people touch the writing, and how much a slip costs.

Which one fits which situation

Some situations point clearly one way.

One person, many small jobs, tight budget. HyperWrite. Three personas on the entry plan is genuinely better value than most of this category, and the toolkit around them does real work.
You want research with citations and a browser extension. HyperWrite. ScriptGrain doesn't pitch itself as a general toolkit.
A brand voice shared across a team or agency. ScriptGrain. A profile you can inspect, edit and score against gives a group something to agree on.
You need to catch drift over months. ScriptGrain. Voice drift is a measurable gap there, not a feeling.
You work in ChatGPT, Claude or Cursor and want the voice to follow you. ScriptGrain, through MCP or the API.
You're unsure what your voice even is. ScriptGrain's Free plan gives one profile and one full analysis, no card.

Sometimes both make sense. If you love HyperWrite's one-click tools and extension but you've been uneasy about consistency, keep it for the quick jobs and use a measured profile as the checkpoint for anything that carries your name or your brand's.

At the end of the day, the question is whether "sounds about right" is enough. For a newsletter you write alone, often it is. For a client deliverable passed through four pairs of hands, it usually isn't.

Using ScriptGrain alongside HyperWrite

Running both is straightforward, because they don't compete for the same slot in your workflow. HyperWrite is where you draft fast. ScriptGrain is where you measure.

A workable pattern: generate a first pass in HyperWrite with a persona, then paste it into ScriptGrain's brand voice consistency checker and read the score and deltas. If it lands in the on-voice band, ship it. If not, the deltas tell you what to fix, and you can run Polish or edit by hand.

If you're moving away from HyperWrite rather than alongside it, start with the writing you've already published, not the examples you pasted into personas. The profile is built from a body of finished work, and three or more pieces totalling about 3,000 words gives a steadier read than a pair of favourites. The writing style analysis tool is a low-effort way to see some of what gets measured before committing to anything. And if your drafts keep sounding generic, the AI cliché checker is worth a pass too.

One practical note on cost. You can see the current ScriptGrain plans on the pricing page, and Free covers one profile, so trying the pairing costs nothing but time. HyperWrite's prices, as read on 2026-09-20, may have moved since, so check them before you budget.

For a wider view of how the category lines up, the AI writing voice tools matrix sets the tools side by side. So does /vs, if you're weighing more than these two.

Where HyperWrite is stronger

Pitches personas directly at the same problem: "Make the AI write like you, not like a robot."

Three personas on the entry plan, where most rivals give one
A broad set of one-click writing tools beyond voice matching
Browser extension that works across the web

Three personas at the entry tier is genuinely better value than most of this category, and if you want a general-purpose writing toolkit rather than a voice engine, HyperWrite covers more ground.

Questions

Is ScriptGrain better than HyperWrite?

They solve the problem differently. HyperWrite matches the style of the examples you paste. ScriptGrain measures 45 named attributes across everything you have published and keeps that profile as an artefact you can inspect, refine with more samples, and re-check: the measurement is the product, not a step inside it. Three personas at the entry tier is genuinely better value than most of this category, and if you want a general-purpose writing toolkit rather than a voice engine, HyperWrite covers more ground.

How much does HyperWrite cost compared with ScriptGrain?

HyperWrite's entry plan is $19.99/mo ($16/mo billed annually), with this free tier: Referenced on the pricing page (limited monthly credits) but the tier's contents were not stated. ScriptGrain starts at £12/month (Writer, 1 voice profile) with a free plan: See your own voice profile and try one draft. No card. HyperWrite figures read from their pricing page on 3 August 2026.

Does HyperWrite learn my writing voice?

Custom Personas built from examples you provide; its style tool asks for two examples and matches their style and format. ScriptGrain instead measures 45 style attributes from writing you have already published, and scores every draft against that measured profile.