Partilhar via


Azure OpenAI

Connects to Azure OpenAI to perform operations on large language models.

This article describes the operations for the Azure OpenAI built-in connector, which is available only for Standard workflows in single-tenant Azure Logic Apps.

Built-in connector settings

In a Standard logic app resource, the application and host settings control various thresholds for performance, throughput, timeout, and so on. For more information, see Edit host and app settings for Standard logic app workflows.

Connector how-to guide

For more information about integrating Azure OpenAI with your workflow in Azure Logic Apps, see Integrate Azure AI services with Standard workflows in Azure Logic Apps.

Authentication

URL and key-based authentication

Authentication type

Parameters

Name Description Type Required Allowed Values
Azure OpenAI endpoint URL The URL for the Azure OpenAI endpoint. string True
Authentication key The authentication key for Azure OpenAI. securestring True

Active Directory OAuth

Active Directory OAuth

Parameters

Name Description Type Required Allowed Values
Azure OpenAI endpoint URL The URL for the Azure OpenAI endpoint. string True
Active Directory OAuth Active Directory OAuth string True
Authority Active Directory authority string False
Tenant Active Directory tenant string True
Credential type Active Directory credential type string False Certificate, Secret
Client ID Active Directory client ID string True
Client secret Active Directory client secret securestring True
Pfx Active Directory pfx securestring True
Password Active Directory password securestring True

Managed identity

Managed identity

Parameters

Name Description Type Required Allowed Values
Azure OpenAI endpoint URL The URL for the Azure OpenAI endpoint. string True
Managed identity Managed identity string True
Managed identity Managed identity string False

Actions

Get an embedding

Executes a GetEmbeddings request for the specified single text input.

Get chat completions

Gets chat completions for the specified text input.

Get completion (Preview)

Gets completion for the specified text prompt.

Get multiple chat completions (Preview)

Get multiple chat completion options for the specified text input.

Get multiple embeddings

Executes a GetEmbeddings request for the specified array of text inputs.

Get an embedding

Executes a GetEmbeddings request for the specified single text input.

Parameters

Name Key Required Type Description
Deployment identifier
deploymentId True string

The deployment or model name.

Single text input
input True string

The single text to convert to an embedding.

Returns

Name Path Type Description
Embedding
embedding string

An array of floats that represent the input's computed embeddings.

Get chat completions

Gets chat completions for the specified text input.

Parameters

Name Key Required Type Description
Deployment identifier
deploymentId True string

The deployment or model name.

Sampling temperature
temperature string

A value used to control the apparent creativity of generated completions. See Azure.AI.Inference.ChatCompletionsOptions.Temperature Property.

messages
messages True string

The deployment or model name.

Nucleus sampling (top_p)
top_p string

A value used to control the apparent creativity of generated completions and an alternative value to Temperature. See ChatCompletionsOptions.NucleusSamplingFactor Property.

Max tokens
max_tokens string

The maximum number of tokens to generate. See ChatCompletionsOptions.MaxTokens Property.

Presence penalty
presence_penalty string

A value that influences the probability of generated tokens appearing based on their existing presence in generated text. See ChatCompletionsOptions.PresencePenalty Property.

Frequency penalty
frequency_penalty string

A value that influences the probability of generated tokens appearing based on their cumulative frequence. See ChatCompletionsOptions.FrequencyPenalty Property.

Returns

Name Path Type Description
Chat completion output role
role string

The chat completion output role.

Chat completion response
content string

The chat completion response.

Chat user
name string

The chat user.

Get completion (Preview)

Gets completion for the specified text prompt.

Parameters

Name Key Required Type Description
Deployment identifier
deploymentId True string

The deployment or model name.

Sampling temperature
temperature string

A value used to control the apparent creativity of generated completions. See Azure.AI.Inference.ChatCompletionsOptions.Temperature Property.

prompts
prompts True string

The deployment or model name.

stopSequences
stopSequences string

The deployment or model name.

Max tokens
max_tokens string

The maximum number of tokens to generate. See ChatCompletionsOptions.MaxTokens Property.

Presence penalty
presence_penalty string

A value that influences the probability of generated tokens appearing based on their existing presence in generated text. See ChatCompletionsOptions.PresencePenalty Property.

Frequency penalty
frequency_penalty string

A value that influences the probability of generated tokens appearing based on their cumulative frequence. See ChatCompletionsOptions.FrequencyPenalty Property.

Returns

Name Path Type Description
Completion response
text string

The completion response.

Get multiple chat completions (Preview)

Get multiple chat completion options for the specified text input.

Parameters

Name Key Required Type Description
Deployment identifier
deploymentId True string

The deployment or model name.

Sampling temperature
temperature string

A value used to control the apparent creativity of generated completions. See Azure.AI.Inference.ChatCompletionsOptions.Temperature Property.

messages
messages True string

The deployment or model name.

Nucleus sampling (top_p)
top_p string

A value used to control the apparent creativity of generated completions and an alternative value to Temperature. See ChatCompletionsOptions.NucleusSamplingFactor Property.

Max tokens
max_tokens string

The maximum number of tokens to generate. See ChatCompletionsOptions.MaxTokens Property.

Choice count (n)
n string

The number of chat responses to generate. See ChatCompletions.Choices Property.

Presence penalty
presence_penalty string

A value that influences the probability of generated tokens appearing based on their existing presence in generated text. See ChatCompletionsOptions.PresencePenalty Property.

Frequency penalty
frequency_penalty string

A value that influences the probability of generated tokens appearing based on their cumulative frequence. See ChatCompletionsOptions.FrequencyPenalty Property.

Returns

Name Path Type Description
Chat message choices
choices string

The array of chat message choices.

Get multiple embeddings

Executes a GetEmbeddings request for the specified array of text inputs.

Parameters

Name Key Required Type Description
Deployment identifier
deploymentId True string

The deployment or model name.

Array input
input True string

The array input to convert to embeddings.

Returns

Name Path Type Description
Embeddings array
embeddings string

An array with arrays of floats representing the input's computed embeddings.