Turn any website into a brand system.
BrandMeld explores the live web and reconstructs the visual and verbal DNA behind a company.
Try
A brand guide for people, and the same facts as data for agents.
Logo
The mark that belongs to the site, from its own header, not the partner logos beside it.
Colour
Primary, supporting and surface colours, weighted by what the page actually paints.
Type
Heading and body faces, weights and the size hierarchy, measured in a real browser.
Voice
Tone, writing habits and a do and don’t list taken from how the copy is written.
Evidence
Every finding keeps the pages it came from and a confidence score.
Four steps, each one checkable.
Nothing is looked up from memory. If the site doesn’t say it, or the browser can’t measure it, it isn’t in the result.
- 01TinyFish FetchReads the live site
Starts at the homepage, follows the links that matter (about, product, pricing, careers) and reads each page. It does not crawl everything.
- 02TinyFish BrowserRenders and measures
Opens the page in a real browser and reads computed styles, so colours and fonts come from what is painted, not from guesses in the markup.
- 03GeminiReads the writing
Interprets tone, positioning and messaging from the copy, and has to quote the site for anything it claims.
- 04ValidatorRemoves what isn’t supported
Claims that can’t be found on the site are dropped, conflicts are flagged, and everything left gets a confidence score.
One stable schema, three endpoints.
The same Brand DNA shape comes back for every company. Ask for the full record, the presentation data, or ready-to-paste context for a prompt.
- POST /api/extract brand DNA
- POST /api/guide guide data
- POST /api/context agent context
{
"schema_version": "1.0",
"brand": { "name", "url", "tagline" },
"colors": { "primary": { "hex", "rgb", "confidence", "evidence": [ … ] }, … },
"typography": { "heading", "body", "weights", "hierarchy" },
"logo": { "primary", "variants", "favicon" },
"voice": { "tone", "dos", "donts", "terminology" },
"messaging": { "tagline", "positioning", "value_propositions", "ctas" },
"design_tokens": { "color.primary": "…", "font.heading": "…" },
"confidence": { "overall", "by_section" },
"validation": { "checks": [ … ] }
}