Sample voice profiles
Three real voice profiles side by side: the narrative, the measured attributes, and content generated in each voice. See exactly what ScriptGrain produces.
Each preview is a real profile built from real writing. The lexical layer is shown in full here; a profile you build yourself carries all 45 reference points across eight layers.
Tech essayist
Measured from 4 samples, 4503 words. Confidence 0.85.
How this writer writes
You write like a technical essayist who reaches for extended metaphors, whether financial, theatrical, or horticultural, to make abstract engineering ideas land. Your paragraphs build patiently toward a punchy, aphoristic final line, and you favour dry humour over overt emotion. You trust the reader to follow a sustained analogy rather than spelling everything out, and you close by circling back to your opening anecdote.
Measured lexical attributes
- Preferred words: abstraction, legibility, maintenance, artefact, narrative
- Filler phrases: The thing is, in practice, in my experience, for the record
- Vocabulary diversity index: 0.6
- Word-length distribution: long 0.15, short 0.45, medium 0.4
- Rare-word rate: 12.5
- Contraction frequency: 10.5
Written in this voice
I once watched a team spend three weeks evaluating a new observability platform while the actual bug sat, undisturbed, in a function nobody had opened since March.
That's not a tooling problem. That's a thinking problem wearing a tooling costume.
The thing is, tools promise legibility. You buy the dashboard, the dashboard shows you the shape of the mess, and for a moment it feels like understanding. But a dashboard is just a mirror with a subscription fee. It reflects the mess; it doesn't resolve it.
In my experience, the teams who ship reliably aren't the ones with the sharpest stack. They're the ones who sit with a problem long enough to see its actual outline before reaching for anything.
So before the next procurement call, try this instead: an hour of silence, a whiteboard, no Slack.
Unhurried thinking has no vendor. That's exactly why it's undervalued.
Newsletter writer
Measured from 4 samples, 4871 words. Confidence 0.83.
How this writer writes
You write conversational essays that build from a small, specific anecdote to a quiet universal point, favouring warm directness over polish. Your sentences swing between long, dash-laced build-ups and short punchy closers, often addressing the reader directly. Dry, self-deprecating humour and honest asides like 'here's the thing' keep the tone intimate rather than lecturing. You trade formal precision for rhythm and narrative pull, closing almost every piece on a terse, memorable line.
Measured lexical attributes
- Preferred words: quietly, shelf, ledger, appetite, properly
- Filler phrases: Here's the thing, Which is to say, here's the thing about, a small confession, for the record
- Vocabulary diversity index: 0.62
- Word-length distribution: long 0.12, short 0.54, medium 0.34
- Rare-word rate: 13
- Contraction frequency: 28
Written in this voice
A team I worked with once bought a £40,000 platform to solve a problem that turned out to be one bloke not talking to another bloke on the next floor.
True story. Nobody wanted to hear it at the time.
Here's the thing about new tools: they feel like progress because they arrive with a receipt. A demo, an onboarding call, a Slack channel full of enthusiasm. You can point at them in a meeting and say, look, we did something.
Unhurried thinking doesn't come with a receipt. It looks like sitting quietly with a problem until you actually understand its shape, rather than its symptoms. Nobody claps for that. There's no vendor to thank.
But most of what breaks in teams isn't a tooling gap. It's a thinking gap, dressed up to look like a tooling gap because that's the version with a budget line.
So before you buy the fix, sit with the problem a while longer.
The ledger doesn't lie, even when the invoice does.
Strategy operator
Measured from 4 samples, 4198 words. Confidence 0.88.
How this writer writes
You write like an operator briefing a board: conclusion first, numbers stacked as proof, risks and next steps labelled in caps. You favour short declarative sentences, numbered structure, and named owners with deadlines over any stylistic flourish. Warmth and humour are absent by design; clarity, authority, and accountability drive every choice.
Measured lexical attributes
- Preferred words: runway, consolidation, attributed, exception, net contribution
- Filler phrases: worth noting, we estimate that, that said
- Vocabulary diversity index: 0.445
- Word-length distribution: long 0.15, short 0.46, medium 0.39
- Rare-word rate: 8.5
- Contraction frequency: 0.375
Written in this voice
Most teams buy a tool to solve a thinking problem.
That never works. We have watched it fail across a dozen consolidation projects. The tool arrives, the licence gets expensed, and the underlying problem sits there untouched for another quarter.
Here is what we have found instead.
First, the teams with the best output are not the ones with the best stack. They are the ones who protect two or three hours a week of unhurried, undistracted thinking. No Slack. No standup. Just the problem, on paper, worked through slowly.
Second, we estimate that most "we need a tool" requests are actually "we haven't sat with this long enough" requests wearing a disguise.
Third, unhurried thinking has a compounding effect that no dashboard replicates. It surfaces the actual constraint, not the symptom.
Therefore: before the next tool purchase, block the thinking time first. Test the constraint. Buy the tool only if the constraint survives contact with a clear head.