Duaer

Integrations

Use Microsoft Foundry Chat Model in Duaer

In Duaer, Microsoft Foundry Chat Model is an AI sub-node. Connect it to the Chat Model input of AI Agent, Basic LLM Chain, or another AI root node.

Connect Microsoft Foundry Chat Model in Duaer

  1. On the Duaer canvas, select the Chat Model input under AI Agent, Basic LLM Chain, or another AI root node and pick Microsoft Foundry Chat Model.
  2. Under Authentication, choose API Key or Azure Entra ID (OAuth2), then pick or create the matching credential (Microsoft Foundry (API Key) or Microsoft Foundry (Entra ID)) in Credential to connect with.
  3. Set Model (Deployment).
  4. If needed, select Add Option for the optional settings listed below.
  5. Run AI Agent. Microsoft Foundry Chat Model runs as part of it, and its calls show under that node in the execution.

Microsoft Foundry Chat Model parameters in Duaer

  • Project: The Microsoft Foundry project that owns the deployment.
  • Model (Deployment): The deployment to use.
  • Model Family: The API family the deployment answers on.

Microsoft Foundry Chat Model options in Duaer

Under Add Option, Microsoft Foundry Chat Model offers Frequency Penalty, Maximum Number of Tokens, Response Format, Presence Penalty, Sampling Temperature, Timeout (Ms), Max Retries, Top P, and Extra Body.

Questions

Where does Microsoft Foundry Chat Model connect in Duaer?

In Duaer, Microsoft Foundry Chat Model connects to the Chat Model input of AI Agent, Basic LLM Chain, or another AI root node. It does not run on its own.

Which credential does Microsoft Foundry Chat Model use in Duaer?

Duaer's Microsoft Foundry Chat Model uses Microsoft Foundry (API Key) or Microsoft Foundry (Entra ID), depending on the method you choose under Authentication.

How do you pick a deployment for Microsoft Foundry Chat Model in Duaer?

A Foundry credential can pick the deployment from the list. A classic Azure OpenAI credential uses id and you type the deployment name. For a deployment that only serves the Responses API, turn on Use Responses API when Model Family is OpenAI.

In this section