November 16, 2026

How to Connect Your Brand Guidelines to Claude and ChatGPT

Add a brand kit to a Claude or ChatGPT Project, or connect your brand guidelines once over MCP so every AI tool reads the current version.

To get Claude or ChatGPT to follow your brand guidelines, give each one a brand kit: your voice, audience, messaging pillars, approved claims, banned words, and a few real examples, kept in a Claude Project, a Claude Skill, or a ChatGPT Project. That works well for one person. For a team, the weak point is upkeep, because copies drift. Connecting your guidelines once over MCP (Model Context Protocol) fixes that: in Marcora, your positioning, brand voice, and writing style live in one Brand Foundation, and Claude or ChatGPT can read the current version. Update a fact once, and every connected tool reads the change. Setup takes about 2 minutes, the setup time listed on both the Claude connect page and the ChatGPT connect page.

What goes in a brand kit?

A brand kit is the short set of material an AI tool needs before it writes a word for you:

  • Voice: how you sound, in a few plain sentences.
  • Audience: who you are writing for, and what they care about.
  • Messaging pillars: the three or four things you want people to remember.
  • Approved claims: what you can say about your product, and the exact wording.
  • Banned words: the phrases you never use.
  • Real examples: three to five pieces you were proud to publish.

Keep the rules short and put them in the instructions. Put the longer guide and the examples in files. Both Claude Projects and ChatGPT Projects work this way: instructions on one side, uploaded files on the other.

How do you set up a brand kit in Claude?

In a Project. Create a project, add your kit documents to its knowledge, and write the rules in the project instructions. Anything in a project's knowledge is used across all of the chats in that project, and the instructions apply to all of them too. A few limits are worth knowing. Project files can be up to 30MB each, and Claude reads text only from most documents, so a diagram inside a Word file is not read. The number of files is not capped, but the total content has to fit in Claude's context window. On paid plans, projects scale beyond that with retrieval (RAG). On Team and Enterprise plans you can share a project with colleagues as "Can view" or "Can edit."

As a Skill. A Skill is a folder of instructions, scripts, and resources that Claude loads when it is relevant to the task. Anthropic lists applying brand style guidelines to documents and presentations as a use for custom skills. Skills need code execution turned on. A skill you upload is not visible to colleagues until you share it. On Team and Enterprise plans, you can share it with specific people, and recipients get your updates the next time they use it.

How do you set up a brand kit in ChatGPT?

In a Project. Select New project, upload your kit as files, and add the rules as project instructions. Project instructions apply only within that project and override your global custom instructions. The number of files depends on your plan: 5 per project on Free, 25 on Go and Plus, and 40 on Edu, Pro, Business, and Enterprise, with 10 files uploaded at a time. You can share a project, and OpenAI's own example for content creation is applying project-specific instructions to keep tone and style consistent across contributors.

What about a custom GPT? If your brand kit is a custom GPT, plan for a change. OpenAI has announced that custom GPTs are scheduled to retire on December 11, 2026, and offers a migration that moves a GPT's instructions and knowledge files into a plugin. New GPTs can no longer be created on personal Free, Go, Plus, or Pro accounts. We compare the options in Claude Projects, ChatGPT Projects, or an MCP Connector: Which Keeps Your Messaging Consistent?.

Where does a brand kit start to strain for a team?

A brand kit in a Project or Skill is a real answer, and for one person in one tool it is often enough. It starts to strain in four places:

  • It lives in one tool. A Claude Project, a Claude Skill, and a ChatGPT Project are three separate places. A team that uses both Claude and ChatGPT ends up keeping more than one copy of its guidelines.
  • Files are copies. A file you upload is the document as it was that day. When the original changes, someone has to remember to replace the copy, in every place you uploaded it.
  • Shared kits still need an owner. A shared Claude Skill is view-only, which keeps it steady. A shared ChatGPT project is described by OpenAI as one that will "evolve over time due to sustained user contributions," which is a strength for collaboration and a risk for a single source of truth.
  • A kit is a document, not a library. A brand guide fits in a file. Product docs, buyer personas, and competitive analysis add up quickly, and in a Claude Project the total has to fit in the context window.

Where does connecting once win?

Connecting replaces the copies with a source. In Marcora, your Brand Foundation (positioning, brand voice, and writing style) is always included, and a Reference Library of product docs, buyer personas, competitive analysis, and messaging is scored for the audience you are writing for. Claude, ChatGPT, Cursor, and any AI tool that supports MCP connections read from the same place. Start a draft in one tool and finish it in another, with its context intact.

That wins when:

  • Your team uses more than one AI tool. You update the source once. Nobody re-uploads anything.
  • You have more than a brand guide. Product facts, personas, and competitive context are kept in the Reference Library, not stuffed into one file.
  • Someone owns the truth. A governed source has an owner. A pile of uploaded copies does not.

If your whole team lives in one AI tool and one short guide, a shared Project or Skill may be all you need.

How do you connect Marcora to Claude?

  1. Open claude.ai/directory/marcora and click Connect. Marcora is listed in Claude's connector directory.
  2. Approve access to finish the sign-in.
  3. Ask Claude "What's my brand foundation?" The first time Claude uses each Marcora tool, it asks you to approve it. Click Always allow.

Prefer to paste a URL? In Claude, go to Customize, then Connectors, click Add, choose Add custom connector, and enter https://mcp.marcora.ai. The full steps are in the Claude connect guide.

Connect Marcora in ChatGPT

Connect Marcora in ChatGPT from the plugin directory. You'll need a Marcora account. Open the Marcora listing and click Install plugin, sign in to Marcora and approve access, then start a new chat and ask "What's my brand foundation?" ChatGPT draws only on the context your team has added to Marcora. The full steps are in the ChatGPT connect guide.

Is there a brand voice connector for ChatGPT?

Yes. Connect Marcora in ChatGPT from the plugin directory. You'll need a Marcora account. Once it is connected, your Brand Foundation and Reference Library are available to ChatGPT, and OpenAI says ChatGPT automatically uses your installed plugins when they are relevant to your request. To call Marcora directly, type @Marcora in the message box. Try: Use Marcora in ChatGPT to tighten this paragraph so it matches our brand voice.

Frequently asked questions

How long does it take to connect Marcora?
About 2 minutes. That is the setup time listed on the Claude connect page and the ChatGPT connect page.

Do I still need a brand kit if I connect Marcora?
Your Brand Foundation does the kit's job: positioning, brand voice, and writing style, always included. If you already have a brand guide, use it to fill in your Brand Foundation, and keep the longer material in the Reference Library.

Should I use a Project, a Skill, or a connector?
Use a Project or a Skill when one person or one team works in one tool with a short, stable guide. Use a connector when you work across tools, keep more than a guide, or need one owner for the facts. You can combine them: Anthropic says MCP connections give Claude access to tools, while skills teach Claude how to use them.

Does this work in other AI tools?
Yes. Marcora works with Claude, ChatGPT, Cursor, and any AI tool or agent that supports MCP connections.

What happens to my custom GPT?
OpenAI has announced that custom GPTs are scheduled to retire on December 11, 2026, and offers a migration to plugins. See our comparison for what moves and what does not.

See it work: the Claude connect guide, the ChatGPT connect guide, or How Marcora assembles context before the first word is written.