Use BlacklistGuard with Claude, ChatGPT and other AI apps

Connect an AI app to your BlacklistGuard account and you can run your email work by asking for it in plain language. Check a domain against the blacklists, clean a list, look at how a campaign did or send one, all from the chat you already have open.

What this is

BlacklistGuard has an MCP server. MCP, short for Model Context Protocol, is the open standard that lets AI apps use outside services on your behalf. Claude, ChatGPT, Hania and a growing number of other apps support it.

Once you connect one of them, the app can see what your account can do and carry out those actions when you ask. You do not copy API keys around or write any code. You sign in to BlacklistGuard once, approve the app, and that is it.

The server address is the same for every app:

https://app.blacklistguard.com/mcp

Things you can ask for

A few requests to give you the idea. You do not need special wording. Ask the way you would ask a colleague.

• "Is mail.example.com on any blacklist right now?"
• "Validate these 40 addresses and tell me which ones are risky."
• "How did last week's newsletter do? Which domains bounced the most?"
• "Show me the DNS records my new sending domain still needs."
• "Add these three people to the Customers list."
• "Build a segment of subscribers who opened something in the last 30 days and tell me how many match."
• "Draft a campaign to the Product updates list with this text, but don't send it yet."
• "How many emails have we sent this month, and how close are we to our limit?"

Before you start

You need two things:

• A BlacklistGuard account on a plan that includes API access. At the time of writing every plan includes it, the Free plan among them. See pricing for what your plan covers.
• An AI app that can connect to a remote MCP server and sign in with OAuth. Claude, ChatGPT and Hania all can. For other apps, see Connect any other MCP app.

Any member of your workspace can connect an app. The app acts as the person who connected it, so it can do what that person can do and nothing more.

Connect Claude

Claude calls these custom connectors. They are available on every Claude plan, with a limit of one custom connector on the free plan.

1. In Claude, open Settings and then Connectors.
2. Choose to add a custom connector.
3. Give it a name, for example BlacklistGuard.
4. Paste the server address, https://app.blacklistguard.com/mcp, and continue.
5. Claude detects that the server uses sign-in. Keep the suggested options and select Add.
6. Claude sends you to BlacklistGuard. Sign in if you are asked to, review the request and select Allow. You are returned to Claude.

To use it in a chat, select the + button next to the message box, open Connectors and switch BlacklistGuard on.

On a Claude Team or Enterprise plan an owner adds the connector once under Organization settings and Connectors. Each member then connects their own BlacklistGuard account from their own Connectors page.

Connect ChatGPT

ChatGPT lets you add your own MCP server as a custom app. What you can do depends on your ChatGPT plan, and in a ChatGPT workspace an admin may have to switch on developer mode before members can add one.

1. In ChatGPT, open Plugins and choose to create a new custom MCP app.
2. Enter a name, for example BlacklistGuard, and a short description.
3. For the server address enter https://app.blacklistguard.com/mcp. Keep the /mcp at the end.
4. For authentication choose OAuth. You do not need to enter a client ID or secret.
5. Read the warning ChatGPT shows about custom servers, confirm, and create the app.
6. ChatGPT sends you to BlacklistGuard. Sign in if you are asked to, review the request and select Allow.

In a new chat, type @ in the message box and pick BlacklistGuard, or choose it from the list of apps.

Connect Hania

In Hania you add BlacklistGuard as an MCP server, and your agents can then use it.

1. In the Hania console, open MCP servers and choose Connect custom.
2. Enter a name, and for the endpoint enter https://app.blacklistguard.com/mcp.
3. For authentication choose OAuth and connect. The server is saved and shown as awaiting authorization.
4. Select Authorize. Hania sends you to BlacklistGuard. Sign in if you are asked to, review the request and select Allow.
5. Back in Hania the server becomes active and its tools are loaded for you. Give them to the agents that should use them.

Hania's own guide has more detail: MCP servers.

Connect any other MCP app

If your app is not listed above, look for a setting called something like "add MCP server", "custom connector" or "remote server", and give it this address:

https://app.blacklistguard.com/mcp

Choose OAuth as the sign-in method if the app asks. There is nothing else to fill in. The app works out the rest by itself.

An app can connect if it meets these three conditions:

• It connects to remote MCP servers over HTTP. Apps that only run local servers on your own machine cannot reach it.
• It signs in with OAuth. BlacklistGuard API keys are not accepted here. They are for the HTTP API.
• It does not need a client secret. Most MCP apps do not.

Claude Code, for example, connects with one command and then asks you to sign in:

claude mcp add --transport http blacklistguard https://app.blacklistguard.com/mcp

The approval screen

Whichever app you use, connecting always ends on the same BlacklistGuard page. It is the normal BlacklistGuard sign-in, so two-factor authentication and Sign in with Google work as usual. Once you are signed in the page shows:

• Which app is asking, with its web address.
• What it will be able to do. There are two levels, described below.
• Which workspace it will work in, and which account you are signed in as.
• Where you go next. This is the app's own address, which you return to after you answer.

Select Allow to connect or Deny to refuse. Either way you are sent back to the app. A request waits for ten minutes. If you take longer, reload the page or start again from the app.

Read access and write access

Write access includes read access. The level is the one the app asks for, and most apps ask for both. You cannot change it on this screen. You can allow the request as it stands or deny it.

Two warnings you may see

• Unverified app. The app registered itself with BlacklistGuard instead of identifying itself by its own web address. That is normal for some apps, Hania among them. Continue only if you started the connection yourself, from an app you trust.
• This app runs on your own computer. You see this for desktop tools and command line tools. Continue only if you started it yourself.

What a connected app can do

A connected app can reach nearly everything you can do in the console. You do not need to learn any commands. The app sees what is available and chooses for you.

Admins and members

An app has the same reach as the person who connected it. Two things are reserved for workspace admins: adding or changing users, and changing API keys. A member's app is not offered those. Everything else is available to every member, exactly as in the console.

What an app cannot do

• Move money. An app cannot top up your balance, change your plan, change auto top-up or add or remove a payment method. It can only read your billing.
• See or create secrets. Passwords, API keys, SMTP passwords, signing secrets and two-factor codes never pass through an app. You create and view those in the console only.
• Do more than you can. Every action goes through the same checks as the console and the API.

Actions are real

Most apps show you what they are about to do and ask before they act. Read that prompt before you agree, especially for sending and deleting. Whether an app asks first is decided by the app, not by BlacklistGuard, so it is worth checking how yours behaves before you give it a large job.

A good habit is to ask for a draft first. "Draft a campaign but don't send it" leaves you a draft to check in the console before anything goes out.

Usage and billing

Connecting an app is free, and so is asking it questions. What counts is the work it does for you. An email sent through an app, an address validated or a blacklist check counts toward your plan's allowance and your prepaid balance in exactly the same way as if you had done it in the console or over the API. Nothing costs more through an app, and nothing costs less.

Your plan's limits apply as well. If you have reached a limit, the app is told so in the same words the console would use.

Limits worth knowing

See and manage your connected apps

Open your Profile in the console. Two tabs cover connected apps.

Connected apps

Every app you have connected, with the access it has, when you connected it and when it was last used. Select Disconnect to end one. The app loses access straight away, and you can connect it again from the app whenever you like.

App activity

A record of what connected apps did: which action ran, when, for whom and whether it worked. You can filter by app and by date. Workspace admins see the activity of every member and can filter by member. Everyone else sees their own.

How long a connection lasts

A connection stays in place for as long as the app keeps using it. If an app goes 30 days without using its connection, the connection ends and you approve the app again next time.

A connection also ends when:

• you disconnect it in Connected apps, or the app disconnects itself;
• you change or reset your password;
• a workspace admin sets a new password for you.

A password change ends all of your connections at once. If you think an app has access it should not have, changing your password is the quickest way to cut everything off.

An app is refused while your user account or your workspace is deactivated, or while your plan does not include API access. If that is put right, the connection works again without a new approval.

Changing your email address, changing your role or turning two-factor authentication on or off does not disconnect anything. A role change takes effect on the app's next action.

What we record

For each action we record which tool ran, who it ran for, which app made the call and from what address, whether it worked, how long it took, and the ids it referred to. If an action fails we keep the error message. We never receive your conversation with the app, and we do not store the other details of a request or its result.

We also record when an app is connected and when it is disconnected. You can see all of this yourself under App activity.

The AI app handles your conversation under its own terms and privacy policy, so check those for the app you use.

Troubleshooting

Still stuck? Contact us and tell us which app you are connecting and what it showed you.