AI Tokens Management

Configure AI tokens to enable secure communication with third-party AI providers.

The AI Tokens view allows you to store and manage all the secret tokens in a secure and unified way across the Emporix Agentic AI. The tokens can be reused when needed across the AI functionalities.

The tokens play a critical role in enabling secure communication with third-party AI services, such as OpenAI for instance, or external MCP servers, and other integrations. Tokens within the Emporix Agentic AI are secured through a multi-faceted approach designed to ensure both convenience and compliance without compromising security.

You might need AI tokens in the following or similar use cases:

• Connecting to OpenAI or other tools with your own enterprise API token for GDPR-compliant production environments

• Granting agent access to external MCP servers without exposing raw credentials

• Managing credentials for other third-party AI services that you want to use in custom agents or integrations

Tokens security

Here's how tokens are secured by design:

• Centralized management and control - AI Tokens view acts as a single point of control. This centralization makes it easier to create, manage, and reuse tokens, preventing them from being scattered across multiple services or configurations.

• Encryption - Every token entered is encrypted immediately before being stored in the database. Once stored, the token value itself is never returned through the API, even when using a GET request. This design protects sensitive information end-to-end and prevents accidental exposure of credentials in logs, dashboards, or integrations.

• Internal use only - Tokens are only used internally by Emporix services when communication with external systems is required. This approach minimizes risk while still allowing smooth interoperability with third-party providers.

• Immutable token values - Once a token value is provided, it is immutable. This means the secret value cannot be changed later. While the token's name can be updated for a more descriptive label, the secret value remains locked for its entire lifetime. This immutability ensures consistency, allowing services dependent on the token to trust that its underlying value has not been altered.

• Secure invalidation process - If you ever need to invalidate a token (for example, due to token compromise or rotation by a third-party provider), the process workflow ensures clarity, control, and the complete retirement of old or insecure tokens.

Managing tokens

In the Management Dashboard, go to the Agentic AI -> AI Tokens to manage the tokens. Each token is represented as a separate card.

Creating a token

1

Choose to create a new token

In the Agentic AI -> AI Tokens, choose Add new token.

2

Provide the token details

Fill in the necessary information about the token:

• Token ID - A token identifier that is used as a reference to the token in other parts of the system by APIs.

• Token Name - A human-readable name of the token, usually containing information about the token purpose.

• Token Value - the secret key to be stored and secured, and later reused. The token value you enter gets already encrypted.

3

Save the token

Select Save to securely store the encrypted token in the database. Further configuration allows only to edit the Token ID or Token Name.

Invalidating a token

For security reasons, the token values cannot be edited. On the occasion you need to invalidate an existing token, follow these steps.

1

Create a new token

Create a new token with the updated value. Follow the steps in the Creating tokens section.

2

Replace usages of the old token

Check where the old token was used (which AI agents or integrations). Edit each of them and replace the old token with the newly added.

3

Remove the outdated token

Once, the old token is replaced, in the Agentic AI -> AI Tokens view, find the old token and choose Remove.

As a result, the token is securely removed and no longer available.