AI Agent Endpoint Target

Last updated
Save as PDF

An AI Agent Endpoint target connects Cisco Workflows to a large language model (LLM) provider, such as OpenAI, Anthropic, or AWS Bedrock. The target stores the provider, region, and model that the AI Prompt and AI Agent activities use.

Before you begin

Gather the following information before creating the target:

The LLM provider to use: AWS Bedrock, Azure OpenAI, OpenAI, Anthropic, Google Gemini, or OpenRouter.
The deployment region for the provider endpoint: AP, EU, or US.
The model identifier to run, which depends on the selected provider.
A matching account key for the provider. See AI Agent Endpoint Credentials.

Provider configuration

The Provider Configuration section holds the required settings. The model list filters by the selected provider, so choose a provider first.

Field	Required	Description
Provider	Yes	The AI provider for the endpoint: AWS Bedrock, Azure OpenAI, OpenAI, Anthropic, Google Gemini, or OpenRouter.
Region	Yes	The deployment region for the provider endpoint: AP, EU, or US.
Model name	Yes	The model that runs the request. The available options change based on the selected provider.

Provider-specific settings

The target shows one settings section based on the selected provider. The following tables describe the fields for each provider. Fields not listed are hidden for that provider.

AWS Bedrock settings

Field	Required	Default	Description
AWS Region	No	us-east-1	The AWS region where Bedrock is available, such as us-east-1 or us-west-2. The Bedrock model must be enabled in the AWS account.

Azure OpenAI settings

Field	Required	Default	Description
Azure Endpoint	Yes	None	The Azure OpenAI resource endpoint, such as https://your-resource.openai.azure.com.
Deployment Name	Yes	None	The deployment name created in Azure OpenAI Studio. The model name must match this deployment.
API Version	No	2024-02-01	The Azure OpenAI application programming interface (API) version.

OpenAI settings

Field	Required	Default	Description
Base URL	No	https://api.openai.com/v1	The OpenAI API base uniform resource locator (URL).
Organization ID	No	None	The OpenAI organization identifier.
Project ID	No	None	The OpenAI project identifier.

Anthropic settings

Field	Required	Default	Description
Base URL	No	https://api.anthropic.com	The Anthropic API base URL.

Gemini settings

Field	Required	Default	Description
Google Cloud Project ID	No	None	The Google Cloud project identifier, used for Vertex AI.

OpenRouter settings

Field	Required	Default	Description
Base URL	No	OpenRouter	The OpenRouter API base URL.

Advanced settings

Field	Required	Default	Description
Action Timeout (seconds)	No	180	The maximum time to wait for an AI operation to complete.

Supported providers and credentials

Each provider requires a matching account key. The following table maps each provider to its credential type.

Provider	Credential type
AWS Bedrock	AWS Credentials
Azure OpenAI	Azure OpenAI API Key
OpenAI	OpenAI API Key
Anthropic	Anthropic API Key
Google Gemini	Google Gemini API Key
OpenRouter	OpenRouter API Key

Create an AI Agent Endpoint target

Perform the following steps to create an AI Agent Endpoint target:

From the left navigation pane, click Automation > Targets.

Click New Target.

From the Target Type drop-down list, choose AI Agent Endpoint.

In the General section, enter a display name and a description for the target.

In the Provider Configuration section, select the provider, region, and model name.

In the provider-specific section, enter the settings for the selected provider.

In the Account Key section, select an existing account key, or click Add New to create one.

Click Save to save the target.

Next steps

Create the matching account key. See AI Agent Endpoint Credentials.
Assign the target to a workflow that uses the AI Prompt or AI Agent activity.
Adjust Action Timeout (seconds) if AI operations need more time to complete.