Branded House
Marketing.Chat
BRANDED HOUSE × MARKETING.CHAT
EXCLUSIVE OFFER · COMMUNITY

Your brand is non-negotiable. Your AI should be too.

Welcome to the end of the Sameness Era. Marketing.Chat is the first AI marketing operating system that turns your brandbook into code - so your brand never loses its voice, distinctiveness or consistency.

  • Persistent Brand DNA across every model (Claude, Gemini, GPT, Grok)
  • Visual DBAs protected with up to 14 simultaneous reference images
  • White-Glove onboarding with our team embedding your brandbook
11,000+PROFESSIONALS60+BRAND SPECIALISTS4AI ENGINES
marketing.chat / brand DNA
Write a LinkedIn post announcing our launch.
BRAND DNA APPLIED · 3 GUARDRAILS ACTIVE

Version aligned with brandbook

LinkedIn · 1 / 1

Today we turn on a light that's been taking shape for months. It's not a "new product". It's a piece of our conviction made operational. We did the hard thing on purpose: said no to the obvious, picked few vectors, defended what matters. Whoever works with us will feel it before reading it.

TONE: confident editorialPALETTE: cobalt + aquaDBA: "conviction made operational"

Every output is shaped by agentic skills that invisibly inject your brandbook rules - before the model writes a single word.

01 / MANIFESTO

We know what keeps you up at night.

We follow the conversations. We know you debate Ehrenberg-Bass theses, the creative emotion measured by System1, and the thin line between Differentiation and Distinctiveness. We know the point isn't "cheap leads" - it's protecting the most valuable asset the company has, when 80% of brand impact still comes from the long term.

We know how it hurts to watch a long-term strategy diluted by quarterly pressure, or to see your Tone of Voice flattened by a generic AI that writes the same way for every company. Pasteurization isn't a synonym for efficiency - it's the shortest path to Sameness.

That's why Marketing.Chat has a single obsession: building technological guardrails for your brand. So that AI stops being a threat to originality and becomes the engine of your consistency.

02 / HOW IT WORKS

From intent to delivery in 3 steps.

No six-month implementation. No prompt-engineer hires. Your brand boards, connects, and the team produces with consistency from the first brief.

  1. 01

    Onboard your brandbook

    Manuals, decks, tone guides, photography repertoire, anti-patterns - upload everything in PDF, DOCX or PPTX. Our isolated vector search absorbs every guideline and keeps it in its own namespace.

  2. 02

    Connect your stack

    GA4, Google Ads, Meta Ads, RD Station and Google Drive plug in via Model Context Protocol. Data is read on demand only - we don't copy your CRM.

  3. 03

    Brief your team

    They ask for "Q3 awareness campaign", "3 versions of the same post on brand" or "a 10-slide deck for the board" - consistency travels invisibly inside every output.

03 / THE FOUR PILLARS

Four guardrails your brand deserves.

Each pillar answers a concrete pain felt by people who have built brands for over a decade. In each card: what it is, why it matters and what you actually get.

BRAND DNA

A voice that survives the model.

WHAT IT IS

Brandbook embedded in an isolated vector namespace per project. Brand Voice, palette, jargon, approved examples and anti-patterns live in their own index with semantic RAG and Relevance Gate. Memory persists across conversations, modes and models.

WHY IT MATTERS

Unlike per-chat prompts or generic "custom GPTs", your brand stays consistent when you switch between Claude, Gemini, GPT and Grok inside the same conversation. No context reset on each model swap.

WHAT YOU GET

Intern, freelancer and CMO produce in the same voice. Tone audits stop becoming review queues. The brandbook's ROI no longer depends on who actually read the PDF.

DISTINCTIVE BRAND ASSETS

Visual Hammer protected by design.

WHAT IT IS

Proprietary Nano Banana 2 visual studio with up to 14 simultaneous reference images, 4K output, 1-click background removal and unlimited PNG/WebP/JPG/PDF exports. Veo 3.1 (Fast and Pro) for HD video clips with audio, up to 8s, ready for reels and campaign teasers.

WHY IT MATTERS

Not generic Midjourney, not pure Sora: generation is anchored in your real DBAs - palette, shapes, photographic repertoire, typography - extracted from the embedded brandbook. The internet's statistical average doesn't beat your identity.

WHAT YOU GET

Visual distinctiveness preserved from print to reel. Your brand is recognized without the logo. Goodbye visual pasteurization and "this doesn't feel like us" rework loops.

SKILLS + 3 MODES

Consistency becomes infrastructure, not goodwill.

WHAT IT IS

Three operating modes (Chat, Copilot, Genius) with modular skills declared in code. A conditional router evaluates intent before each message and invisibly injects your brand guidelines. Genius mode runs parallel agents for multi-artifact deliveries.

WHY IT MATTERS

Instead of a 4,000-token mega-prompt distributed via Slack, your brand rules become infrastructure: applied automatically, without the team having to remember every anti-pattern in the brandbook.

WHAT YOU GET

Junior team produces at senior standard. New hire onboarding stops taking 30 days. You don't wake up to off-brand work - because off-brand doesn't ship.

REASON + EMOTION

Specialists who read GA4 before they write.

WHAT IT IS

60+ vertical specialists - Branded Content, Brand Strategy, Editorial Copy, Growth, Paid Media, SEO, Controller, Videomaker - living in one interface. Native MCP integrations with GA4, Google Ads (incl. MCC for agencies), Meta Ads, RD Station CRM/Marketing and Google Drive. Encrypted tokens, on-demand reads.

WHY IT MATTERS

Cold data + human repertoire in one surface. The Branded Content specialist writes knowing the channel's real CTR. The Paid Media analyst briefs copy without changing tabs. End of the "Excel on one side, Word on the other" routine.

WHAT YOU GET

Brand decisions backed by real data, but with the right soul. Committee meetings get shorter because the material arrives with a diagnosis - not a question.

04 / THE FULL STACK

All of this underneath, in a single subscription.

When we say "marketing operating system", it's not a metaphor. Each piece below is part of the product you subscribe to - no add-on, no hidden upsell.

SPECIALISTS

60+ verticals

Branded Content, Brand Strategy, Editorial Copy, Growth, Paid Media, SEO, Controller, Videomaker. Personas tuned to real brand problems - not generic chatbot.

VISUAL STUDIO

Nano Banana 2

Proprietary engine up to 4K. Up to 14 references per generation, 1-click background removal, unlimited PNG/WebP/JPG/PDF exports on paid plans.

VIDEO

Veo 3.1

HD clips with audio up to 8s. Fast (~53 credits) and Pro (~142 credits) modes. Useful for teasers, campaign cuts and narrative BTS.

MODES

Chat · Copilot · Genius

Chat for fast responses across 8 models in one conversation. Copilot for guided step-by-step briefings. Genius as an autonomous factory with parallel agents.

MEMORY

Persistent Voice & KB

Brand Voice, Brand DNA, anti-patterns and approved examples configured once. Applied to every chat, image, video and artifact - across every model.

MULTI-MODEL

Claude · Gemini · GPT · Grok

Claude Opus 4.6, Sonnet 4.5, Haiku 4.5; Gemini 3.1 Pro (1M tokens) and Flash Lite; GPT-4o and 4.1; Grok. Switch models mid-conversation without losing context.

MCP INTEGRATIONS

GA4 · Ads · Meta · RD · Drive

Native connectors via Model Context Protocol. Includes Google Ads MCC for agencies and end-to-end RD Station CRM. Encrypted tokens, on-demand reads.

PROJECTS + KB

Per-brand isolated namespace

Each customer in their own vector index (Vertex AI) and separate MongoDB Atlas. Up to 300MB of KB on Scale+. Zero cross-tenant leakage, isolation by design.

05 / IN PRACTICE

From strategy to deck - without switching tabs.

Four real deliverables brand builders run inside Marketing.Chat. Not features - the work you already do, now with the infrastructure underneath.

01INTEGRATED BRIEFING

Campaign brief crossing media + brand + data

The Branded Content specialist writes the narrative while Paid Media briefs the channel - both with GA4 and Meta Ads open, reading real data via MCP. Final output: a single briefing with objectives, KPIs and campaign copy already aligned with the brandbook tone.

02BOARD MEETING

Board presentation, in editable Slides

Genius mode builds the deck (10–13 slides) with relevant photos from Nano Banana 2, GA4 charts pulled live, and editorial copy passed through the Brand Strategy specialist. Saves to Drive already converted to native Google Slides - 10 to 15 seconds from click to editing.

03GEO AUDIT

How Claude, Gemini, GPT and Grok answer about you

GEO Monitor runs 14 parallel queries across the four generative engines and returns a 6D score, competitive comparison (up to 5 brands) and action plan. The first audit is free - useful to see where your brand is being erased from AI's answer.

04FULL CAMPAIGN

14 visual variations + 3 video cuts + editorial copy

Genius mode fires in parallel: Nano Banana 2 generates 14 variations aligned to your DBAs, Veo 3.1 delivers 3 teaser cuts, and the Editorial Copy specialist closes the copy. All within the brandbook, ready to publish - in a single execution.

06 / DATA

Your brandbook is yours. Period.

When you upload a 200-page brand manual, you're trusting us with the company's most sensitive asset. We treat it accordingly - with the infrastructure enterprise InfoSec teams expect before signing.

  • Zero Data Training

    No company data - including the files you upload to projects - is used to train the public AI models (OpenAI, Google, Anthropic, xAI). What belongs to the company stays in the company. Contractual guarantee with each provider.

  • Private Google Cloud buckets

    All uploads (pipeline spreadsheets, PDFs, images, videos) are stored with AES-256 encryption at rest, TLS 1.2+ in transit and exclusive access via temporary URLs with 1-hour expiration.

  • Tenant isolation

    On the Teams plan, your environment is shielded: you have absolute control over who enters, what they access and how data flows. Each workspace in its own namespace on MongoDB Atlas and dedicated vector index on Vertex AI - zero cross-tenant leakage.

  • LGPD + audit-ready

    Designated DPO, configurable retention per workspace, exportable access logs in CSV, revocable OAuth on every connector. Ready for the legal team's review without exceptions.

07 / IN PRODUCTION

Who already trusts their brandbook on the platform.

From hyper-growth startups to century-old brands. Professionals from over 11,000 companies - many of them in the same conversations you are - already use our infrastructure to orchestrate Claude, Gemini, GPT and Grok without giving up their identity.

11,000+
ACTIVE COMPANIES
40+
COUNTRIES
8
AI MODELS
08 / TRANSPARENCY

Not for everyone - and that's fine.

Where Marketing.Chat shines

  • Brand with a structured brandbook: documented tone of voice, DBAs and visual rules
  • Team with process maturity - briefings, approval flows and brand governance
  • Active operation across at least 2 channels (paid media + content) with regular volume

Where it doesn't make sense yet

If your brand is still under construction - no tone manual, no defined DBAs, no long-term strategy - the technology delivers less than it could. Before onboarding the AI, it's worth onboarding the brand: our Branding and Strategy specialists help you get there.

Talk to a brand specialist
Branded House
Marketing.Chat
09 / EXCLUSIVE OFFER

A deal shaped for the people who build brands.

As a thank you to the Branded House community and the high-level exchanges you keep alive - Byron Sharp vs. Mark Ritson, Ehrenberg-Bass, System1, the end of the funnel - we built something special for the group's members.

  • Special pricing across all plans

    Exclusive discount disclosed inside the group, valid for any plan. Single coupon, no rate hikes during the subscription term.

  • White-Glove Onboarding

    Our team onboards your brandbook with you. Voice setup, visual palette, DBA library and vertical specialist configuration - perfect from Day 1, without you having to become a prompt engineer.

  • Direct line to the product team

    Your critiques, suggestions and requests turn into roadmap. You speak directly to who builds - not to a scripted SDR.

No catches, no asterisks. Applies to all plans. Automatic invoice via Conta Azul, in BRL, tied to the registered tax ID. Cancel anytime - no fee, no friction.

10 / FAQ

Before you hit the button.

Ten questions that tend to show up among people who take brand seriously - and the people who own governance, data and legal on the other side.

Is Brand DNA just a fancier prompt? How does the AI actually "learn" my brand?
Not a prompt. Each brand lives in an isolated vector namespace on Vertex AI with semantic RAG and Relevance Gate. When you upload brandbook, manuals, decks and approved examples, the AI doesn't paste a mega-prompt into every chat - it queries that index at runtime and injects guidelines invisibly via agentic skills. Memory persists across conversations, modes (Chat / Copilot / Genius) and models (Claude, Gemini, GPT, Grok). Switching models mid-conversation does not reset the brand context.
Can I lock forbidden words and anti-patterns the AI must never use?
Yes. Brand Voice has a dedicated anti-patterns field - terms that must never appear (AI clichés, forbidden jargon, competitor vocabulary, words your brand avoids by positioning). You also declare approved examples ("this SOUNDS like the brand") and rejected ones ("this doesn't sound"). Agentic skills validate every output - if an anti-pattern slipped through, the output is rejected and regenerated before it reaches your screen.
How do you prevent the AI from inventing a number, claim or assertion about the brand?
A universal anti-hallucination rule applied across the four models: nothing gets said without a traceable source. When the AI cites data (e.g. "4.2% CTR in May"), it must point to the origin - MCP tool call to GA4, dispatcher briefing, your message or uploaded file. Without a source, it answers "I don't have that data, want me to fetch it?" instead of guessing. Real case before the fix: the AI invented 62 fictional deals totaling R$ 327k to fill an empty list. That's closed at the platform level.
Does it work for multi-client agencies? Each client with isolated brandbook and data?
Yes - this is a primary case for agencies and holdings. Each client becomes a Project with its own namespace: isolated KB (up to 300MB on Scale+), dedicated Brand Voice, separate vector index on Vertex AI. The Google Ads connector includes MCC for reading all linked accounts without swapping credentials. On the Teams plan, you manage granular permissions for who on the team can access which client. Zero risk of Client A's copy leaking into Client B's briefing.
Does visual generation respect my exact HEX palette?
Yes. The palette declared in Brand Voice is injected as a generation constraint, not a suggestion. You can also upload up to 14 reference images per generation in Nano Banana 2 - real product photos, frames from previous campaigns, mood boards. The engine anchors output in those real visual assets (not the statistical average of the internet), preserving palette, shapes and photographic repertoire. For critical cases, you can lock variations and export only the ones approved in the review flow.
Can junior team members publish directly? Is there an approval flow?
Depends on the plan. On Teams, you configure roles (Editor, Approver, Admin), publication gates and an auditable history of who did what - exportable in CSV for audit. Junior produces, senior approves, the record stays. On individual plans (Pro, Scale, Max), control is per user. Important: agentic skills infrastructure already blocks anti-patterns automatically, so even without a formal flow, off-brand work doesn't ship - the first approval layer happens before the output reaches the screen.
Where are the servers physically located? Can I sign a custom DPA?
Model processing runs on Anthropic, Google Cloud, OpenAI and xAI infrastructure - mostly in the US, with standard contractual clauses and LGPD art. 33 coverage. Structured data (KB, brandbook, conversations, integrations) lives on MongoDB Atlas and Google Cloud Storage. For Enterprise, we offer Brazil-restricted processing options where viable and a custom DPA (Data Processing Agreement) signed by our legal team - including extra clauses for demanding InfoSec teams. We're in the SOC 2 process.
I already use ChatGPT Enterprise + Custom GPTs on the team. Why adopt Marketing.Chat?
ChatGPT Enterprise, even with Custom GPTs, is a chat window: each GPT runs isolated, doesn't talk to the others, doesn't pull GA4 / Meta / RD at runtime, has no persistent cross-model memory, doesn't include Nano Banana 2 (4K visuals) or Veo 3.1 (video) or 60+ vertical specialists or the Copilot and Genius modes. Marketing.Chat is the operating system above: you don't replace ChatGPT, you add brand infrastructure, live data integrations and a multi-artifact delivery factory. Many customers keep ChatGPT for personal use and Marketing.Chat for brand work.
What's included in White-Glove Onboarding?
Structured session with our team (~4-6h spread across two weeks) covering: Brand Voice extraction from your existing brandbook, palette and DBA library setup in Nano Banana 2, selection of vertical specialists relevant to your industry, integration setup (GA4, Google Ads with MCC if you're an agency, Meta Ads, RD Station, Drive) with minimum scopes, and training for the day-to-day operators. Output: Marketing.Chat in production with your brand applied from Day 1, without you having to become a prompt engineer.
Which plan does the coupon work on? Can I cancel anytime? Do you issue invoices?
The Branded House community coupon works on all paid plans (Pro, Scale, Max and Teams), monthly or annual, with no rate hikes during the term. Cancel directly from your account at any time - monthly ends at the close of the 30-day cycle, annual follows Brazilian consumer law (refund within 7 days, no substantial use). Automatic invoice via Conta Azul, in BRL, tied to your CPF or CNPJ. Business invoicing available on Teams and Enterprise plans. Card payment via Stripe (we don't store card numbers) or boleto / PIX.

Your brand deserves infrastructure that matches it.

Claim the community offer, onboard your brandbook with our team, and watch AI become a guardrail - not a threat.