Brand Atlas

Check

Check before you ship.

Unsupported copy can be allowed, rewritten to a line you approved, or held. Brand Match scores what Create makes. It runs when you create here, or when a request goes through the gateway, API or MCP — not a watch on every chat. Not legal advice.

  • Brand Match
  • Claims Check
  • Gateway
  • API
  • MCP

How it works

The policy you kept. Applied before it ships.

  1. 01

    Start from Claims

    Approved lines, evidence and packs already live in the Atlas. Check has nothing to apply until they do.

  2. 02

    Copy comes in

    From Create, a paste, the gateway, the API or MCP. Not from watching every chat on the team.

  3. 03

    Allow, rewrite, or hold

    Unsupported copy can pass, be rewritten to a line you approved, or wait for a human.

  4. 04

    Brand Match on Create

    Voice, type and presentation are scored on what Create makes. It is not a live monitor of every AI.

Why it matters

A brand firewall. Not a hope.

01

The policy, applied

Claims Check can allow a line, rewrite it to wording you kept, or hold it. The model does not get a second vote.

02

Brand Match on Create

What Create makes is scored against voice, type and presentation. It is not a live score of every AI on the team.

03

Where you send it

Gateway, API and MCP. Create. It is not a universal output firewall, and it does not watch every model the team opens.

What it actually does

Two jobs. Same Atlas.

Claims Check applies the policy. Brand Match scores Create. Neither is a live monitor of every AI.

01

Claims Check

Incoming copy is read against the pack you activated. Invented substantiation does not ship.

“Clinically proven to improve focus.”

“A scent tool for a focused state.”

  • BLOCK
  • REWRITE
  • HOLD

02

Brand Match

On Create: voice, type, product presentation. Advisory on the asset — not a watch on every chat.

Brand Match

  • Voice ✓
  • Typography ✓
  • Product presentation ✓

Create

Why it exists

Not a live monitor of every AI.

Check runs when you create in Brand Atlas, or when a request goes through the gateway, API or MCP. It does not watch every chat. Not legal advice.

Hoping it behaves

Every model is on its own.

  • A stronger claim writes itself in.
  • Nothing runs unless someone remembers to look.
  • Every tool keeps a different version of the policy.

Check

The policy, applied.

  • Allow, rewrite to an approved line, or hold.
  • Runs on Create, or through gateway, API and MCP.
  • Not a universal firewall. Not legal advice.

Build your Atlas

Check needs a policy to apply.

Brand, Products and Claims are the source. Check is how that policy is applied — not a second brand book.

Put it to work

Connect and Create.

Connect feeds every AI the same Atlas. Create is optional. Check still runs before you ship.

Check it before it ships.

Questions

How is this different from Claims?

Claims is the policy in your Atlas — approved lines, evidence and packs. Check is how that policy is applied: Brand Match on Create, and Claims Check via gateway, API or MCP before you ship. Same source. Different job.

Does it watch every AI the team uses?

No. It is not a live monitor of every model or chat. Check runs when you create in Brand Atlas, or when a request goes through the gateway, API or MCP.

Is this legal advice?

No. It is a brand firewall, not a legal database. It enforces the claims, evidence and packs you put in. It does not replace counsel. Not legal advice.

Where does Check actually run?

On Create for Brand Match. For Claims Check: when you paste copy, or when traffic goes through the gateway, the API or MCP. It does not intercept tools you never connected.

What is Brand Match?

A read on what Create makes — voice, type and product presentation — against the Brand Book. It is advisory on that asset. It is not a live score of every AI on the team.