An API connection authenticates and scopes access to the Titan AI APIs. You create and manage connections in the Titan platform settings, where an administrator can view every connection for the organization, create new ones, and revoke access. This article explains how to create a connection, connect an AI tool, and manage existing connections.
For more information on our API reference, including usage limits and pagination, see our developer portal.
Types of connections
Titan supports two types of connections, and you create both through the same workflow. A public API connection authenticates requests to the Titan REST API, so you can use it for integrations, dashboards, and automation that call Titan AI directly. An MCP connection authenticates an AI assistant that connects to Titan through the Model Context Protocol (MCP), so you use it when you want to query Titan AI from your own AI tool.
Create a connection
- To create a connection, go to Connectors > API and select Create Credential.
- Enter the following details to generate the credentials you use to authenticate:
- Label: Give the connection a distinct, recognizable name so you can tell your connections apart later.
-
Scope: Choose which Titan capabilities the connection can use. The scope limits the connection to only the actions you select.
- Read security findings: Grants read-only access to supply chain findings.
- Flag security findings: Marks findings as resolved, dismissed, or in progress.
- Select Generate to create a client ID and secret. The secret appears only once, so copy it right away and store it in a password manager or another secure location.
After you generate a connection, Titan provides example code snippets for both the REST API and MCP to help you implement it. If you lose the secret, you regenerate the connection rather than recovering the original value.
Connect an AI tool with MCP
For an MCP connection, Titan AI provides an MCP URL that you use to set up the connection in your AI tool:
- From the Credentials Management page, select Connect AI Client, then copy the provided MCP server URL.
- Paste it into your AI client's option for adding an MCP server.
- Authorize the connection when your AI client prompts you in the browser.
Once connected, you can start to use the MCP in your tool. For an example of how to use the connector in Claude, see this page and video.
Manage existing connections
From the connections list, you can review and maintain every connection for your organization. Each connection shows details such as its name, type, scope, and status.
From this list, you can:
- Check the last time the connection was used.
- Edit the connection to adjust the scope or the roles allowed to use it.
- Delete a connection if it is no longer needed.
You can also see which users have connected to Titan AI through a commercial AI tool, so you can monitor MCP usage across your organization.