Brandefense MCP. Put Brandefense Inside Your AI Agent.

MCP replaces custom glue code with one standard connection. Incidents, exposure, threat intelligence and reports sit behind one server address. The main areas:

One Question, End To End.

Follow a single prompt from your agent to the platform and back.

1. 00:00 Ask You ask in plain language"investigate incident #4021"No query syntax, no dashboard. Type it the way you would ask a colleague.
2. 00:01 Authorize The server checks who is askingSign-in check · your role · your organizationThe agent works with your role.
3. 00:02 Call Reads the incidentincident_detail(4021)The agent calls the platform directly.
4. 00:03 Call Checks the indicatorintelligences_list(ioc: "185.220.101.0/24")The threat intelligence behind the indicator.
5. 00:04 Call Looks at historyincidents_list(ioc: "185.220.101.0/24")Earlier incidents from the same source.
6. 00:05 Correlate Connects the evidencematch: known Tor exit range, seen in earlier incidentsInfrastructure, history and impact, cross-referenced.
7. 00:06 Answer An answer you can act onRoot cause: credential stuffing from Tor exitsRoot cause, next step and the tool calls behind it.

Sample run on acme.com. Times are illustrative.

Same access as the platform. No separate account.

1. 01

   Enable Access

   An organization manager turns connections on once. Everyone in scope can then connect.

2. 02

   Install from the Claude directory

   Open the Brandefense listing in the Claude directory and add it. Sign in with your Brandefense account and approve access.

   Add to Claude ↗

3. 03

   Verify It Works

   Start a new chat with Brandefense switched on and ask “Who am I on Brandefense?” You should see your name, organization and role.

   PromptCopy

   Who am I on Brandefense?

Same access as the platform. No separate account.

1. 01

   Enable Access

   An organization manager turns connections on once. Everyone in scope can then connect.

2. 02

   Open the app in ChatGPT

   Open the Brandefense app in ChatGPT and add it. Sign in with your Brandefense account and approve access.

   Open in ChatGPT ↗

3. 03

   Verify It Works

   Start a new chat with Brandefense switched on and ask “Who am I on Brandefense?” You should see your name, organization and role.

   PromptCopy

   Who am I on Brandefense?

Other ways in: Brandefense In ChatGPT.

Examples of what an agent can call. Read-only tools run directly. Tools that change data ask you to confirm in the chat.

Sample output on acme.com.

Type these the way you would ask a colleague. Names such as INC-4021 and acme.com are sample data.

Who am I on Brandefense?

What needs my attention today?

investigate incident #4021

Summarize incident INC-4021 and who is assigned.

Which phishing sites are still active against my brand?

Start a takedown for this phishing site.

Example calls

What the agent runs behind a question. You do not type these.

incident_detail("INC-4021", format: "summary")

search_blackhole_domain("acme.com")

phishing_site_detail(id)
start_phishing_takedown(id)

• What is MCP?

  Model Context Protocol, an open standard that lets AI agents call outside tools and read outside data in a common way.

• Which clients are supported?

  Brandefense is listed in the Claude directory and as an app in ChatGPT. For other clients that support MCP, your partner can confirm what Brandefense supports today.

• Do I need a separate account?

  No. It gives the same access as the platform. Sign in with your Brandefense account.

• Do I have to copy an API token?

  No. You sign in with your Brandefense account and approve access, so there is no token to copy and store.

• Can the agent see more than I can?

  No. The agent works with your own Brandefense role. Tools can only do what your own account can do.

• How do I check that it works?

  Start a new chat with Brandefense switched on and ask “Who am I on Brandefense?” You should see your name, organization and role.

• “You don’t have permission to use the MCP server”

  An organization manager has not turned connections on, or your team is not in scope. Ask a manager to enable it, then try again.

• A tool answers “Unauthorized” or “Forbidden”

  Your role in the platform does not allow that action. Tools can only do what your own account can do.

• Do some tools need confirmation?

  Tools that change data or start investigations ask you to confirm in the chat before they run.

• How do I disconnect?

  Remove the connection in your client, or ask a manager to switch connections off.