Brand editor, app instructions, and a smarter Checker

Edit your Brand Twin in place, give individual apps their own checking instructions, and get sharper analysis with fewer noisy issues.

A big quality-of-life update across the Brand Twin, Checker and Watcher.

Edit your Brand Twin like a document

The Brand Twin overview is now fully editable in place — no more re-running the wizard to fix a tagline or rename a colour.

  • Colors with editable hex values and names ("Lime", "Onyx", "Saffron"). Add, rename, recolour, delete.
  • Tagline, mission, vision, positioning, elevator pitch, target audience — inline edit with one Save.
  • Do's & Don'ts can be edited freely, plus you can now add channel-specific overrides: a different Don't list for Instagram than for LinkedIn, separate guidance for email or your website.
  • The same colour editor is now available during the initial curation step.

Apps as configurable workers

Every Checker and Watcher app now has its own App context & instructions section. Tell the engine exactly what to focus on for that workspace:

"Check that all body copy is at least 14pt, headings use the brand serif, and the copyright reflects the current year. Compare layout against the attached template."

Upload a reference template (PDF or image) and the engine compares structure, hierarchy, and copy patterns directly against it. App instructions are treated as the highest-priority checks — they're evaluated even when the brand spec doesn't mention that aspect.

Mood image and voice generate themselves

The first time you finalise a Brand Twin, the mood image and brand voice now kick off automatically. No more clicking "Generate" twice. Re-finalising an existing Twin won't re-spend on assets that are already there.

Sharper analysis, fewer repeat issues

  • Issue lists are de-duplicated server-side: five variations of "inconsistent icon style" now collapse to one well-prioritised finding.
  • Each dimension caps at 5 distinct issues — the worst ones rise to the top, the noise stays out.
  • A "Repeated patterns" summary on the asset detail page shows the count of grouped findings at a glance.
  • The Checker asset details now include a top-level summary card with severity counts, repeats, and a list of dimensions that came back compliant.
  • Filter chips on the issues tab let you slice by Visual Identity, Voice & Tone, Messaging, Quality, or Brand Presence.

Better LinkedIn, Facebook, and Instagram scanning

The social scrapers were rewired so a login wall can never quietly poison the brand analysis. When a profile can't be reached, the scan now fails loudly instead of pretending the login page is your content.

Fixes & polish

  • Full rescan of an existing Brand Twin now lands on a proper progress view (it used to bounce you to the wizard's first step). When the rescan completes you go straight back into curation with the fresh content.
  • Asset checker Content tab shows a clean "binary asset" placeholder for PDFs and images instead of the raw [object ArrayBuffer] text.
  • Copyright check is no longer fooled by the current year — the engine knows today's date.
  • A handful of [brandName] placeholders that occasionally leaked into generated copy are now caught and replaced server-side.
  • Watcher now spells out exactly what it scans (the URL you enter, not links from it) so you know whether you need to add separate entries for sub-domains.

Coming soon

  • Deeper visual identity extraction from uploaded brand guidelines (typography hierarchy, layout patterns, iconography)
  • Issue feedback that feeds back into your Brand Twin
  • Tooltip explanations on every metric in the dashboards