IntelBase / MCP Server
IntelBase, inside your AI assistant
Ask a question in ChatGPT, Claude or Cursor. Your assistant runs the email and username lookups and reads the results with you.
Included with the Max plan. Works with ChatGPT, Claude, Cursor, VS Code and more.
Ask it like you'd ask a colleague
"This signup looks fake. Does the email have any real accounts behind it, and how old are they?"
"Where else is this username used, and do the profiles look like the same person?"
"Which data breaches has our support address appeared in, and is it in any stealer logs?"
What your assistant can do
Email lookups
See the accounts registered to an address and the details each site shows.
Username lookups
Find where a username is in use, with the details each profile shows.
Your saved lookups
Ask about something you already looked up, without using another lookup.
Breach sources
See which breaches an address appears in, by name and date, and how many records there are.
Set it up in a couple of minutes
All you need is an IntelBase account on the Max plan. Choose how you'd like to do it.
Let your AI walk you through it
Copy this prompt and paste it into your assistant. It reads our setup guide and takes you through it step by step.
Set it up yourself
Pick your assistant and follow the steps. They use this server URL.
https://api.intelbase.is/mcp- Open Plugins in ChatGPT, select the plus button, then Add custom MCP server.
- Name it IntelBase, paste the server URL under Connection, and choose OAuth for authentication.
- Confirm the warning, create the plugin, then sign in to IntelBase and approve the connection.
- Install the plugin. Then, in a chat, type @ and select IntelBase.
OpenAI limits custom MCP servers to some ChatGPT plans. If you don't see the option, check its documentation for yours.
Menu names change between versions, so check your assistant's documentation if a step looks different. Using another assistant? It works with any client that supports remote MCP servers over Streamable HTTP with OAuth.
Questions
Which plan do I need?
The MCP server is included with the Max plan. Each lookup uses one of your daily lookups, and your plan's rate limit applies. Reading your saved lookups doesn't use one.
How long does a lookup take?
About 20 seconds. You can ask your assistant for a quick check or a thorough search, which sets the wait between 5 and 30 seconds. A shorter wait is faster, but sites that haven't answered by then are left out. If that happens, your assistant can tell you and offer to run it again with a longer wait, which uses another lookup.
Can my assistant read my past lookups?
Yes. Ask about something you already looked up and it can find and read the saved result, along with the date it ran. It only sees your own lookups, not your team's, and breach and stealer log records get the same limits as a new lookup.
Can the assistant see passwords from breaches or stealer logs?
No. For breaches, the assistant gets the source's name and date and the number of records, but never what the records contain. For stealer logs, it gets whether records exist, how many and of what type. To see the records your plan includes, use the IntelBase dashboard.
Do I need an API key or IP whitelisting?
No. You sign in to IntelBase and approve the connection. Your API key isn't shared with the assistant, and assistants don't call from a fixed IP address, so your IP whitelist doesn't apply to this connection.
How do I disconnect an assistant?
Go to your account page and choose Disconnect under Connected AI apps. Its access ends straight away.
Where do lookup results go?
Results are sent to the assistant you connected and handled under its terms, so check that they suit the data you work with.
Does my assistant need to support anything special?
It needs to support remote MCP servers over Streamable HTTP with OAuth sign-in. If yours does, use the server URL above.