Brand voice consistency checker

By Jack Stovell · published 2026-08-07

ScriptGrain scores how closely a draft matches your measured brand voice, so "does this sound like us" stops being an argument and becomes a number. Shared profiles mean several writers generate against the same standard instead of each defaulting to their own instincts.

One baseline, consistent output, whoever happens to be writing today.

Check your own content first. The free off-voice check below takes the piece that sounds most like you and up to three others, pasted or by URL, and scores each one against it with the measured differences behind the number. No signup, and nothing you paste is stored.

Start free: no card, one profile, one full analysis, and one free draft, on the house.

Why voice drifts (and why nobody notices until it's bad)

Here's the thing: drift isn't one big mistake. It's death by a thousand small ones. Three writers, two AI tools, six months, and suddenly your blog reads like four different companies took turns writing it. Nobody signed off on that. It just happened, slowly, the way these things do.

The usual fix is a style guide nobody reads and a reviewer going "hmm, feels off" in a comment thread. That's not a system. That's a vibe check.

1. A measured baseline, not a mood board

ScriptGrain runs a two-pass stylometric analysis across your real copy and pulls out 45 attributes: sentence rhythm, contraction rate, hedging, humour register, the lot. That becomes your baseline. Not an adjective list. A number.

2. A voice-match score on every draft

Every piece of AI-generated copy gets scored against that baseline before it reaches a reviewer. High score, ship it. Low score, you know exactly where it slipped, tone, structure, vocabulary, whatever. No more "this feels wrong" meetings. Just a number and a reason.

3. One baseline, every writer

Profiles are shared across the team, so every draft answers to the same 45 measured attributes. Same voice, whoever happens to be writing today. And to be fair, that's most of the value right there: one standard that outlives whichever tool is trendy this quarter.

4. The tie-breaker document

Disagreements about brand voice usually come down to two people with two opinions and no referee. The profile ends that.

5. Built for teams, not just one writer

Studio and Agency plans give whole teams shared profiles, shared scoring, shared standards. No more one writer's taste quietly becoming "the brand."

Plans that scale with your team

  1. Free: one profile, one full analysis, no card required.
  2. Writer, £12/month: one profile.
  3. Operator, £29/month: three profiles.
  4. Studio, £99/month: ten profiles, for teams juggling multiple brands or clients.
  5. Agency, £299/month: unlimited profiles, ten team seats, per-member credit caps so usage stays sane.

Young category, vague promises, everyone claiming they've "cracked brand voice." Fair enough. But a promise isn't a score.

Get your baseline free. No card, one profile, five minutes.

Then watch drift become a number your team can actually check, not a feeling somebody has to defend in a meeting.

Questions

Is the free off-voice check the same scoring as the paid product?

Same engine, thinner baseline. The check measures every piece with the code the product uses on real profiles (sentence rhythm, punctuation habits, pronoun mix, and a small judged set such as formality and argument structure), then treats your main piece as the baseline. A full profile is built from everything you have written and adds vocabulary, signature phrases and the rest of the 45 attributes, so its scores are steadier and it can generate in the voice, not only measure it. Up to three pieces per check, 120 words minimum each, a few checks an hour per visitor.

What does a Voice Match score measure, and what does it leave out?

It measures stylistic alignment: how closely the draft's rhythm, syntax, punctuation and other measured attributes match the profile. That is all. It does not measure quality, factual accuracy, originality, or whether the idea behind the piece is any good. A high score means it sounds like you. It says nothing about whether what you are saying is right, or worth saying.

How do several writers share one voice baseline, and which plans include team seats?

Everyone generates against the same shared profile instead of writing from memory or a style guide nobody reads. Team seats come with Studio, which includes three, and Agency, which includes ten. Per-member credit caps come with the team wallet, and style guides are available on Studio and Agency.

What happens when a draft scores low?

A polish loop revises the draft toward the profile using the measured divergences, then scores it again, so you are not left rewriting by hand or guessing what went wrong. The supporting analysis also shows which features aligned and which moved away from the target.

Can I compare two brand voices against each other?

Yes, on a ten-axis view plus a categorical difference table. It needs two completed profiles, so in practice it starts on the Operator plan. It is useful for comparing one brand's voice against another, or seeing how far apart two writers actually are before standardising them onto one baseline.