---
name: brand-voice
description: "Build a source-derived writing voice profile from real posts, docs, or site copy, then reuse it across content and outreach so every draft sounds like the same person instead of generic AI copy. Use before drafting anything user-facing when voice consistency matters: articles, launch notes, social posts, outbound messages. Falls back to Selwyn's own default voice (direct, plain-word, sourced) when no other source material is given, not a generic one."
---

# Brand Voice

Derive a voice from real source material and reuse it, or fall back to my own default voice, sourced from my anti-slop rules, when there's nothing to derive from.

## Phases

Work through these in order.

### 1. Gather

Collect 5 to 20 real samples (posts, articles, launch notes, outbound messages, site copy), preferring recent material unless told older writing is more canonical.

Read: [writing-without-ai-slop](https://selwynuy.dev/d/writing-without-ai-slop.md)

### 2. Extract

Pull the signature, not adjectives: rhythm, compression, how claims get made, what the source never does.

Read: [writing-without-ai-slop](https://selwynuy.dev/d/writing-without-ai-slop.md)

### 3. Write

Produce the structured voice profile. If no source material exists, use the default voice and say so plainly instead of silently substituting it.

Read: [writing-without-ai-slop](https://selwynuy.dev/d/writing-without-ai-slop.md)

## Bundled files

- `references/voice-profile-schema.md` (reference): The structured VOICE PROFILE format a derived voice should follow.

---

Part of Brand Voice from Selwyn Uy's Next.js Handbook. Full page: https://selwynuy.dev/skills/brand-voice