Skip to main content
Voice synthesis is the method that makes a ContentHero draft different from a generic “write me a caption” prompt. ContentHero grounds your agent in your real, performing context, and your agent drafts in your voice, emulating what already works for you. The output is a draft you approve, in your own voice. The rule: grounded voice-synthesis is the feature; generic ungrounded copy is the anti-pattern. If the grounding is missing, the skill asks for it or proceeds without claiming the result is on-brand. It does not fill the gap with generic copywriting.

​
The three grounding sources

Pull all three before drafting. Each answers a different question.
1

Brand voice: how you sound

The brand kit’s voice profile, audience, positioning, and content strategy, plus the sections where banned words and tone guardrails live. This is your intended voice.
2

What performs in your niche: proven patterns

Your top outliers, ranked by performance. Shallow at the list level, so mine the best few for depth.
3

What performs for you: your track record

Your own brand-account top posts and your past captions. The truest model of your voice, and the strongest signal of the three.

​
The list-then-get mining pattern

Outliers come back shallow (title and metrics). The deep fields you draft from come from a second call.
  1. List outliers (or read your own top content) and pick the few most relevant, highest-performing items.
  2. Get the full content of each chosen item for the transcript, description, hashtags, and keywords. These are what you extract patterns from.
Do not mine a hook from the list view. Pick from the list, then get the detail.

​
Extracting pattern primitives

From the deep content of the chosen high performers, extract the durable, transferable patterns, not the literal words:
  • Hook archetype: how the opening earns attention (question, bold claim, pattern interrupt, number, contrarian take).
  • Structure: the shape (problem then payoff, list, story arc, demo then result).
  • Pacing and length: short and punchy versus a longer build.
  • CTA style: how they close and ask for the action.
  • Hashtag and keyword strategy: the volume and specificity that travels in your niche.
Two or three high performers reveal the shape. Emulate the shape in your voice and topic.

​
When the brand voice and your top posts disagree

Your actual high-performing posts win. They are evidence; the stated profile is intent.

​
The draft and approval

Your agent applies the proven patterns to this post’s topic, in your voice, for your audience, and matches the destination platform’s norms. Then it presents the draft, with a brief note on its reasoning, and gets your approval. It iterates on feedback and never proceeds to produce or publish on an unapproved draft.

​
When grounding is thin

The skill always tells you which grounding a draft stands on. With no brand kit, it asks for tone direction or says the draft is not yet brand-grounded. With no outliers, it leans on your past posts; with no past posts, on the niche outliers. That transparency is part of the value: you know the draft emulates real, performing context, not a generic template.