Your brand in Claude, ChatGPT and n8n, with a brand score

The connector URL:

https://branddock.app/mcp

Your assistant never publishes on its own. You approve.

Agents write posts, emails and articles effortlessly these days. But your brand is not in the model: it lives in people’s heads, in a PDF nobody opens, in the instincts of that one colleague who rewrites everything. So you get fluent, generic output, and “does this sound like us?” stays a gut-feel conversation.

Branddock answers that question with a number. Your brand DNA (voice, personas, products, competitors, do's and don'ts) becomes the context of your agent, and anything you submit for a check gets an F-VAL score from 0 to 100 with concrete findings. You approve before anything goes live.

The brand layer: context in, validation on top

Two things make an agent brand-safe.

Context in

One tool call and your agent has your complete brand context as system context: the same layered context stack Branddock itself injects into every AI call.

Validation on top

Have a text, from any source, scored with F-VAL: a composite score from 0 to 100, a threshold check and concrete findings per category. Before publishing, not after.

What your assistant can do with it

Your assistant gets the Branddock pipeline as tools, in three groups: knowing your brand, making it on-brand and checking. Image generation runs through a direct API key, not through the connector.

Know

Free, never costs credits

get_brand_context
get_brand_md
get_design_md
list_brands
list_personas
list_products
list_competitors
search_knowledge
get_deliverable_content

Make

Credits only on generation

generate_on_brand
rewrite_on_brand
generate_long_form_seo
generate_web_page
generate_campaign_strategy

Check

Free, never costs credits

score_against_brand
get_strategy_status
get_seo_status

One brand, two open files

The same scan that feeds your agent also produces the open files of the standard: BRAND.md says who your brand is (voice, audience, guardrails) and DESIGN.md how it looks (colors, typography, sizing). Your agent reads both live, through get_brand_md and get_design_md; the free generator creates them from any website.

Try the free generator

Connect in a few minutes

Same connector URL, three routes.

In Claude

1.Customize (or Settings) → Connectors → add connector
2.Paste the connector URL and confirm
3.Log in with your Branddock account and your brand comes along

In ChatGPT

1.Turn on Developer mode in Settings
2.Connectors → add → paste the same connector URL
3.Sign in with your Branddock account and you are set

Your own stack (API key)

1.Create an API key in Settings → API & Connectors
2.Use the key as a Bearer token on the same URL
3.Unlock extra tools the connector doesn't carry, such as generate_image and import_brand_data

What can your agent do once you consent? Reading your brand and scoring texts is free. Generating costs credits and always lands in your Branddock library. The agent never publishes on its own: you approve.

First prompts to try

Right after connecting: this is how you see what the brand layer does, in one minute.

Fetch my brand context and summarise how my brand sounds.

Write a LinkedIn post about [topic] for persona [name] and score it with the brand check.

Rewrite this email in our brand voice: [paste your text]

For developers

The same brand layer, machine-to-machine. Authentication through brand-locked API keys or OAuth for agents. Generating costs the same credits as in the platform. A key belongs to one workspace, and the credits come from the balance of the organization that workspace belongs to. For agencies: with one key per client workspace, each key works on that client's brand; the credits come from your agency's shared pool.

With the public REST API, the MCP node or webhooks you build the brand check into n8n or your own automation. Webhooks are HMAC-signed.

28 days free · 300 credits · no credit card

Any AI can write. Branddock shows whether it's your brand.