Use BrandViz from Claude, ChatGPT & Cursor
BrandViz runs an MCP server, so you can use it from inside the AI assistant you already work in. Ask Claude how ChatGPT describes your brand this month, have Cursor pull the pages AI assistants cite most, or file next month’s content tasks from a conversation. Answers come from your own reports, not from the assistant’s guesswork.
One connection reaches every brand your BrandViz account can open. Your assistant signs in as you, so there’s no API key to copy or store.
What you need
- A BrandViz account with at least one brand.
- An assistant that supports remote MCP servers with sign-in: Claude (web and desktop), Claude Code, Cursor or ChatGPT. Other MCP clients that support OAuth work too.
The server URL
Every app asks for this address. The first time it’s used, it sends you to BrandViz to sign in and approve the connection.
https://mcp.brandviz.ai/mcpConnect your assistant
Claude
- In Claude (web or desktop), open Customize, then Connectors.
- Choose Add custom connector, name it BrandViz and paste the server URL.
- Press Connect and sign in to BrandViz when asked. Start a new chat to use it.
Claude Code
- Run this in your terminal, then /mcp inside Claude Code and pick brandviz to sign in.
claude mcp add --transport http brandviz https://mcp.brandviz.ai/mcpCursor
- Add to Cursor installs it in one click, or paste this into ~/.cursor/mcp.json. Cursor asks you to sign in the first time it is used.
{
"mcpServers": {
"brandviz": {
"url": "https://mcp.brandviz.ai/mcp"
}
}
}ChatGPT
- In ChatGPT settings, open Apps & Connectors and turn on developer mode under Advanced.
- Create a connector named BrandViz, paste the server URL and choose OAuth.
- Sign in to BrandViz when asked.
Once connected, the app shows up under Settings → AI assistants in BrandViz, with when it was last used.
Things to ask
Start with an open question. The assistant calls overview first to see which brands you have and what data each one holds. Then it drills down from there.
- “How visible is my brand in AI answers this month, and what changed since the last report?”
- “Which competitors does ChatGPT recommend instead of us, and for which questions?”
- “Which of our pages get cited most, and which AI crawlers have been reading the site?”
- “What does the latest site audit say we should fix first?”
- “Add a question about pricing for the UK market, then show me what we’re tracking there.”
- “Draft three content tasks for the gaps in early-stage questions, but don’t run them.”
Name the brand by its slug, the short name in your BrandViz address (for example acme in /b/acme). Your assistant’s approval prompt shows the arguments of each call, and a slug tells you which brand a change is for at a glance.
What a connected app can do
These are the same permissions the sign-in screen shows you before you approve.
- Read the brands this account can already open. Visibility scores, citations, AI traffic, audits, the work queue, the connected repo’s branches and what a job left in the connected CMS. No other account’s brands. (The staff-only funnel read is refused by name to everyone else.) Tools:
overview,visibility,citations,ai_traffic,workstream,audit,job_changes,cms_entry,branches,funnel - Change what a brand measures. Its questions, markets, competitor regions, alias names, brand profile and brand lessons. Takes effect on the next report and the next job. Tools:
brand_profile,questions,regions,competitor_region,aliases,lessons - Record progress on recommended actions. Tick a step, change a status or priority, add a note — the same writes the Actions page makes, and it can undo them. Tools:
actions - File content tasks, and approve or dismiss them. Filing one runs nothing; it waits for a human on the Tasks board. Approving starts the work. Dismissing has no undo. Uses your monthly allowance. Tools:
tasks - Start a content job. Writes a draft to your repo as a pull request, or to your CMS as a draft, or rebuilds the Blog Style Profile. Never publishes. Uses your monthly allowance. Tools:
run_job,blog_style
What it can’t do
- Publish anything — a job’s output always stops at a pull request or a draft
- Reach a brand this account cannot already open
- See or change billing, plan limits, team members, or a stored integration credential
- Delete a brand, a report, or a question’s history — archiving is the app’s own soft delete
When your assistant adds a content task, it lands on your Tasks board as pending and nothing runs. Work only starts when it’s approved, on the board or by an explicit approve call. That is also the point where your monthly allowance is spent. Responses that spend allowance or run AI calls say so, so you can see it in the conversation.
Approvals in your assistant
Every tool carries a standard MCP hint saying whether it only reads, changes settings, or does something that can’t be undone. Clients use these hints to decide when to ask you. Most let you allow read-only tools once and keep asking for the rest. Anything that needs a person in BrandViz comes back with a direct link to the right screen.
Disconnecting
Remove the BrandViz connector in the app that holds it, and that app can no longer reach your brands. Your other connections and your BrandViz account are unaffected.
Troubleshooting
- The sign-in window doesn’t open, or closes without connecting. Check that pop-ups are allowed for your assistant, then try again. You can only connect with an account that can already sign in to BrandViz.
- A brand you expected is missing. The connector sees exactly the brands your account can open in BrandViz. Ask the brand’s owner to invite you.
- A tool says a screen isn’t on your plan. Some features, like content tasks and jobs, come with the Engine. The assistant can still read your visibility data on every plan.
- Nothing happens after connecting in Claude. Start a new chat. Connectors load when a conversation begins.
Privacy and support
The connector reads and writes only the brands your account can open, under the same rules as the BrandViz app. See our privacy policy and terms. For help, email hello@brandviz.ai.