AI Tools · Brand Voice · Personal Branding · Published · 8 min read

How to Use One Brand Definition Across ChatGPT, Claude and Gemini

Write your brand definition once and reuse it in ChatGPT, Claude and Gemini: a master file, each app's context features, and when MCP helps.

To use one brand definition across ChatGPT, Claude and Gemini, write it once as a plain text or Markdown file that you own, then load that same file into each app's context feature: custom instructions or a project in ChatGPT, a project in Claude, and a Gem or skill in Gemini. Keep the master copy outside all three apps and update the app copies from it.

A portable brand definition is a single document, kept outside any one AI app, that states who you are, what you stand for and how you speak in a form any assistant can read.

This guide covers what each app offers, the tools that hold brand context for you, and a setup you can build today without buying anything. App features were checked against official help pages on November 8, 2026. They change often, so check the linked pages for your plan.

→ Jump to: Why Context Does Not Travel | What Each App Offers | Tools That Hold Context | Build Your Own Setup | What MCP Adds

Why Brand Context Does Not Travel

ChatGPT, Claude and Gemini each keep instructions, files and memory inside their own accounts. Nothing you set up in one app is shared with the others. If you explain your positioning in three places, you end up with three versions, and they drift as you edit one and forget the others.

The fix is not a better prompt. It is a source of truth: one document that defines the brand, with every app copy derived from it.

What Each App Offers

ChatGPT

Custom instructions are for preferences you want ChatGPT to follow across chats. They suit a short brand summary and voice rules.

Projects keep related chats, files, instructions and sources together. A project gives its chats access to the same uploaded files, project instructions and connected sources, which suits the full brand document. Start a new chat from inside the project so it uses that context.

Claude

Claude Projects are workspaces with their own chat histories and knowledge bases. You can upload documents to a project's knowledge base and set project instructions. Projects are available on free accounts, up to five projects.

Claude also supports custom connectors using remote MCP on Free, Pro, Max, Ultra, Team and Enterprise plans. Free users are limited to one custom connector.

Gemini

Gems let you save a name, instructions and optional knowledge files to customize Gemini's responses.

Google is moving Gems to skills, so check the current status in Gemini. If you set up a brand Gem today, keep its instructions and files in your master folder so you can rebuild it as a skill.

The same rule applies to all three apps: the app holds a copy, and you hold the original. For a side-by-side look at app access, see the ChatGPT, Claude and Gemini comparison.

Tools That Hold Brand Context for You

If you would rather not maintain copies by hand, three kinds of tools store brand context and pass it on.

Inside a marketing platform

Jasper stores brand voices, style guides, audiences and knowledge in its Jasper IQ layer. Its MCP server gives external AI agents read-only access to that context and to Jasper content creation, with quickstart guides for Claude and ChatGPT. This suits teams that already produce content in Jasper.

Machine-readable brand files

BrandCodex turns a brand identity into structured files, such as a brand manifest, color and typography tokens, a voice and tone file and a rules file, based on an open specification it calls MRBS. Brands can be published to a public registry, and AI agents can access them through MCP. Its examples center on brand systems, including colors and typography.

Markolé builds a brand strategy and visual identity for entrepreneurs and small teams, then offers two ways to share it: an MCP connection that agents such as Claude Code or ChatGPT Codex can read from, and a JSON-LD export. Markolé notes that an export is a copy, not a live connection, and has to be exported again after changes. Markolé is in early access.

BrandKernel

BrandKernel excavates the personal brand of founders and solopreneurs in a guided AI dialogue and encodes it into one reusable brand kernel. The kernel covers 11 layers, from who you are and what you stand for to how you speak and why you exist, with up to ~250 documented fields.

It exports as Web, PDF, Markdown or JSON. The Markdown can go into a Claude or ChatGPT project like any other document; the JSON is structured data for tools. The compiler turns the kernel into task-specific prompts. API and MCP access are available for live connections. The 11-layer method explains how the dialogue works, and what a brand kernel is defines the format.

Build a Portable Brand Definition Without Buying Anything

You can set this up in an afternoon with tools you already have.

1. Write the master file. Use a plain text or Markdown file in a folder you control. Start with this structure and fill each section with your own words:

# Brand definition: [your name]
Version: [date]

## Who I am
One paragraph: what you do, for whom, and your background in facts.

## What I stand for
3 to 5 beliefs about your field, each with one sentence of reasoning.

## Who I serve
The reader or client, the situation they are in, what they need to decide.

## Positioning
Why this audience might choose you over the alternatives they actually consider.

## Voice
Rules, not adjectives: sentence length, words you use, words you avoid.
Two short passages of your own approved writing as examples.

## Claims
Claims you can support, with the source.
Claims to never make (results, guarantees, credentials you do not hold).

2. Make a short version. Condense the file into a summary of a few lines for custom instructions, which suit short preferences. Keep the full file for projects and Gems or skills.

3. Load the same file everywhere. Upload the full file to a ChatGPT project, a Claude project and a Gemini Gem or skill. Add one project instruction in each: "Use the attached brand definition for every draft. Do not invent claims that are not listed in it."

4. Test with one prompt. Ask each app for the same small task, such as the opening paragraph of your services page. Check the drafts against the file: audience, claims, voice. The brand voice exercise helps if the voice section is still vague.

5. Update the master first. When something changes, edit the master file, change the version date, and replace the copy in each app. Never edit an app copy directly.

What MCP Adds

MCP, the Model Context Protocol, is an open protocol that standardizes how applications provide context and tools to AI models. With an MCP connection, an assistant reads brand context from a live source when it needs it, instead of relying on a file you uploaded weeks ago.

That solves the copy problem but not the definition problem. A live connection to a vague brand still returns a vague brand. MCP also depends on the app, plan and permissions you have, and it does not guarantee that a model follows what it reads. Start with a clear master file; add a connection when keeping copies in sync becomes the bottleneck.

Frequently Asked Questions

Can I define my brand once and reuse it across ChatGPT and Claude?

Yes. Keep one master brand file and load it into a ChatGPT project and a Claude project, or, where your app and plan support MCP connections, connect to a live source instead. Update the master first, then the copies.

What are portable brand guidelines?

Portable brand guidelines are a brand definition kept in an open format, such as Markdown or JSON, outside any single app, so any AI assistant or person can use the same version. A PDF can be read by people but is harder for tools to reuse in parts.

Should I use custom instructions or a project for my brand voice?

Use custom instructions for a short summary you want in every chat, and a project for the full brand document and task-specific work. You can use both, fed from the same master file.

What happens to my Gemini Gems?

Google is moving Gems to skills, so check the current status in Gemini. Keep your instructions and files in your own master folder either way.

Does MCP make AI follow my brand voice?

No. MCP gives an assistant access to context and tools. The output still depends on the model, the task and the quality of the brand definition, so review drafts before you publish them.

Next Step

Write the master file first, even a rough one. If you want a guided way to build it, see how BrandKernel compares with using ChatGPT on its own, read what a brand kernel is, or follow the 11-layer method.

Key Takeaways

  • Each AI app stores its own context. A brand you explain in ChatGPT does not reach Claude or Gemini.
  • Keep one master brand file outside the apps, and treat every app copy as a copy.
  • ChatGPT, Claude and Gemini each have a place for reusable instructions and files.
  • MCP lets compatible assistants read brand context from a live source instead of a pasted copy.
  • No setup guarantees that a model follows your brand. Review outputs against the master file.

Related articles

All articles

Your brand identity isn't invented. It's buried. Let's excavate it.