Every AI tool writes in the same confident, anonymous voice. Yours doesn't have to. ScriptGrain doesn't guess your style from a prompt: it measures 45 stylometric reference points from writing you've already done, holds every draft to those constraints, and scores each one against your profile. Miss the voice? A polish loop revises toward you, not generic-good. See your full profile free, no card. UK-built, UK English as standard, API + MCP server included.
Hey, I'm Jack. Solo builder, UK based, and this comment is the demo.
I wrote it with ScriptGrain, using a voice profile measured from my own writing, then ran it through the product's polish loop before posting. Not "write like me" typed into ChatGPT. An actual profile: 45 stylometric reference points covering sentence rhythm, contraction rate, punctuation habits, how I open a paragraph, the words I reach for without noticing.
That's the problem I built this to fix. Every AI tool writes in the same confident, anonymous voice. Custom instructions capture maybe five things you think are true about your writing. They don't capture the habits you don't know you have.
ScriptGrain measures those from writing you've already done, generates new drafts constrained to them, and scores every draft against your profile. If a draft misses, the polish loop revises toward you rather than toward generic-good.
The free plan is a real taster: your complete profile, all 45 reference points, plus one draft. No card required.
Product Hunt folk: PRODUCTHUNT20 at checkout gets 20% off any plan for 12 months.
Built in the UK. UK English is a default setting here, not an autocorrect fight.
Now, tell me where this breaks. Where does it sound off? What would make you not trust the score? I'd rather hear the blunt version than the polite one.