Anatomy of a Brand DNA
The fields Brandflux captures from a brand URL, what is editable, and which fields drive every variation downstream.
A Brand DNA is the source of truth Brandflux anchors every variation to. Five field groups, all editable, most extracted from a single URL.
Palette
Primary, secondary, and up to four accents. Hex values pulled from the live site, ranked by visual prominence on the homepage and product pages. Each color carries a usage note (e.g., “primary CTA, hover state”) that guides the layout engine. You can add or rename colors. Edits propagate to existing campaigns at the next regenerate.
Typography
Headline, body, and mono families with weights and sizes. If the brand uses an italic serif for emphasis, Brandflux flags it during extract review and asks you to confirm. Fallback chains are computed but the headline family is what drives the editorial feel.
Voice
Three to five voice samples (one or two paragraphs each) that capture the formal and informal registers. Add samples by pasting copy from the brand’s site, an existing brand book, or a press release. The Composer reads voice samples at every generate to keep words on-brand.
Imagery
Reference images you mark as “good” (the brand uses these) and “bad” (the brand avoids these). Three of each is the practical minimum. Imagery references shape the look of every generated background, illustration, or photo composite.
Mandatories and do-not-use
Logos, legal lines, banned phrases, banned colors. Anything the brand requires on every asset or refuses to ship goes here. The Composer enforces these as hard constraints, not suggestions.
If a DNA does not feel right after the extract, see Fixing extractions for the extract-or-extend mechanic.