Verified against Claude · 2026-07-11
Get Claude to write in your actual voice, not generic AI voice
A style-matching prompt that feeds Claude several of your own writing samples and asks it to extract and name the concrete patterns before writing anything new, instead of vaguely imitating a vibe.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Below are three or more samples of writing in the voice I want you to match. Before writing anything new, first describe the voice back to me in concrete, checkable terms — not adjectives like "friendly" or "professional," but specific patterns: typical sentence length, how paragraphs open, where contractions appear, what gets a whole sentence versus a fragment, how bluntly disagreement or bad news is stated, and any recurring words or structures. WRITING SAMPLES [Sample 1: a Slack update to the team] [Sample 2: an email to a client] [Sample 3: a paragraph from a blog post] After you describe the pattern, write the following in that exact voice, not a generic approximation of it: NEW PIECE TO WRITE A short LinkedIn post, about 120 words, announcing that we shipped a feature customers have been asking for. CONSTRAINTS - Do not default to more formal or more polished phrasing than the samples actually use, even if it would read better in isolation — matching the samples’ actual register is the job, not improving on it. - If the brief requires a length or format none of the samples demonstrate, say so and make your best-supported guess at how the voice would scale, rather than silently reverting to a generic default. - This one can be slightly more celebratory than the samples, since it is an announcement, not a status update.
Customize the highlighted detailsoptional — the prompt above already works
Why this works
Asking Claude to write 'in my voice' with no examples forces it to guess at a generic idea of a personal voice, which reliably regresses to a smoothed, slightly-too-polished AI default; providing several real samples turns this into a few-shot pattern-matching task, which is a much better-supported capability than voice description from a one-line adjective. The forced intermediate step — describe the pattern back in concrete, checkable terms before writing anything — matters mechanically because it makes the extraction visible and correctable: if Claude describes the voice as 'warm and enthusiastic' when what is actually distinctive is that sentences run long and hedge with 'I think,' you can see that miss and correct it before a single word of new content gets written, rather than discovering the mismatch only in the final draft. The explicit instruction not to polish upward exists because Claude's default writing style is measurably more formal and more evenly hedged than most people's actual working voice, and left unconstrained it will quietly correct 'imperfections' in the sample voice — sentence fragments, blunt disagreement, a recurring pet phrase — treating them as things to smooth over rather than the actual signal being matched.
Verified against
Claude Sonnet 4.6 · 2026-07-11
Changelog
- 2026-07-11 — Initial publish, verified against Claude Sonnet 4.6.
Building this for real?
This is a free starting point. If you'd rather have what Scult builds built and running for your business, that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
