OpenAI

Overview

This module adds support for selected Open AI models.

Maven Coordinates

In addition to the Helidon integration with LangChain4j core dependencies, you must add the following:

<dependency>
    <groupId>io.helidon.integrations.langchain4j.providers</groupId>
    <artifactId>helidon-integrations-langchain4j-providers-open-ai</artifactId>
</dependency>

Components

OpenAiChatModel

To automatically create and add OpenAiChatModel to the service registry add the following lines to application.yaml:

langchain4j:
  providers:
    open-ai:
      api-key: "demo"

  models:
    openai-chat-model:
      provider: open-ai
      model-name: "gpt-4o-mini"

If enabled is set to false, the configuration is ignored, and the component is not created.

Full list of configuration properties:

Key	Type	Description
api-key	string	Required. The API key used to authenticate requests to the OpenAI API.
base-url	string	The base URL for the OpenAI API. If not present, the default value supplied from LangChain4j is used.
custom-headers	Map<string, string>	A map containing custom headers.
enabled	boolean	If set to false, OpenAI model will not be available even if configured.
frequency-penalty	double	The frequency penalty, between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model’s likelihood to repeat the same line.
log-requests	boolean	Whether to log API requests.
log-responses	boolean	Whether to log API responses.
logit-bias	Map<string, int>	LogitBias adjusts the likelihood of specific tokens appearing in a model’s response. A map of token IDs to bias values (-100 to 100). Positive values increase the chance of the token, while negative values reduce it, allowing fine control over token preferences in the output.
max-completion-tokens	int	The maximum number of tokens allowed for the model’s response.
max-retries	integer	The maximum number of retries for failed API requests.
max-tokens	int	The maximum number of tokens to generate in the completion.
model-name	string	The model name to use (e.g., "gpt-3.5-turbo").
organization-id	string	The ID of the organization for API requests.
parallel-tool-calls	boolean	Whether to allow parallel calls to tools.
presence-penalty	double	The presence penalty, between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, encouraging the model to use new words.
response-format	string	The format in which the model should return the response.
seed	int	The seed for the random number generator used by the model.
stop	string[]	The list of sequences where the API will stop generating further tokens.
strict-json-schema	boolean	Whether to enforce a strict JSON schema for the model’s output.
strict-tools	boolean	Whether to enforce strict validation of tools used by the model.
temperature	double	The sampling temperature to use, between 0 and 2. Higher values make the output more random, while lower values make it more focused and deterministic.
timeout	duration	The timeout setting for API requests. See here for the format.
top-p	double	The nucleus sampling value, where the model considers the results of the tokens with top_p probability mass.
user	string	The user ID associated with the API requests.

Proxy

The framework attempts to resolve a proxy using the following strategy:

1. Check for a named service: Look in the service registry for a service of type java.net.Proxy with the name open-ai.chat-model.
2. Fallback to another named service: If the first service is not found, search for a java.net.Proxy service with the name open-ai.
3. No proxy: If neither service is found, the framework will not use a proxy.

Tokenizer

The framework attempts to resolve a tokenizer using the following strategy:

1. Check for a named service: Look in the service registry for a service of type dev.langchain4j.model.Tokenizer with the name open-ai.chat-model.
2. Fallback to another named service: If the first service is not found, search for a dev.langchain4j.model.Tokenizer service with the name open-ai.
3. Default Tokenizer: If neither service is found, the framework will use the default tokenizer provided by LangChain4j.

OpenAiEmbeddingModel

To automatically create and add OpenAiEmbeddingModel to the service registry add the following lines to application.yaml:

langchain4j:
  providers:
    open-ai:
      api-key: "demo"

  models:
    openai-embedding-model:
      provider: open-ai
      model-name: "text-embedding-3-small"

If enabled is set to false, the configuration is ignored, and the component is not created.

Full list of configuration properties:

Key	Type	Description
api-key	string	Required. The API key used to authenticate requests to the OpenAI API.
base-url	string	The base URL for the OpenAI API. If not present, the default value supplied from LangChain4j is used.
custom-headers	Map<string, string>	A map containing custom headers.
dimensions	int	The dimensionality of the embeddings generated by the model.
enabled	boolean	If set to false, OpenAI model will not be available even if configured.
log-requests	boolean	Whether to log API requests.
log-responses	boolean	Whether to log API responses.
max-retries	integer	The maximum number of retries for failed API requests.
model-name	string	The model name to use.
organization-id	string	The ID of the organization for API requests.
timeout	duration	The timeout setting for API requests. See here for the format.
user	string	The user ID associated with the API requests.

Proxy

The framework attempts to resolve a proxy using the following strategy:

1. Check for a named service: Look in the service registry for a service of type java.net.Proxy with the name open-ai.embedding-model.
2. Fallback to another named service: If the first service is not found, search for a java.net.Proxy service with the name open-ai.
3. No proxy: If neither service is found, the framework will not use a proxy.

Tokenizer

The framework attempts to resolve a tokenizer using the following strategy:

1. Check for a named service: Look in the service registry for a service of type dev.langchain4j.model.Tokenizer with the name open-ai.embedding-model.
2. Fallback to another named service: If the first service is not found, search for a dev.langchain4j.model.Tokenizer service with the name open-ai.
3. Default Tokenizer: If neither service is found, the framework will use the default tokenizer provided by LangChain4j.

OpenAiImageModel

To automatically create and add OpenAiImageModel to the service registry add the following lines to application.yaml:

langchain4j:
  providers:
    open-ai:
      api-key: "demo"

  models:
    openai-image-model:
      provider: open-ai
      model-name: "gpt-image-1"

If enabled is set to false, the configuration is ignored, and the component is not created.

Full list of configuration properties:

Key	Type	Description
api-key	string	Required. The API key used to authenticate requests to the OpenAI API.
base-url	string	The base URL for the OpenAI API. If not present, the default value supplied from LangChain4j is used.
custom-headers	Map<string, string>	A map containing custom headers.
enabled	boolean	If set to false, the OpenAI model will not be available even if configured.
log-requests	boolean	Whether to log API requests.
log-responses	boolean	Whether to log API responses.
max-retries	integer	The maximum number of retries for failed API requests.
model-name	string	The model name to use.
organization-id	string	The ID of the organization for API requests.
persist-to	Path	The path or location where the generated images should be persisted.
quality	string	The quality of the generated images.
response-format	string	The format of the response.
size	string	The desired size of the generated images.
style	string	The style of the generated images.
timeout	duration	The timeout setting for API requests. See here for the format.
user	string	The unique identifier for the user making the request.
with-persisting	boolean	The flag to indicate whether to persist the generated images.

Proxy

The framework attempts to resolve a proxy using the following strategy:

1. Check for a named service: Look in the service registry for a service of type java.net.Proxy with the name open-ai.image-model.
2. Fallback to another named service: If the first service is not found, search for a java.net.Proxy service with the name open-ai.
3. No proxy: If neither service is found, the framework will not use a proxy.

OpenAiLanguageModel

To automatically create and add OpenAiLanguageModel to the service registry add the following lines to application.yaml:

langchain4j:
  providers:
    open-ai:
      api-key: "demo"

  models:
    openai-language-model:
      provider: open-ai
      model-name: "gpt-4o-mini"

If enabled is set to false, the configuration is ignored, and the component is not created.

Full list of configuration properties:

Key	Type	Description
api-key	string	Required. The API key used to authenticate requests to the OpenAI API.
base-url	string	The base URL for the OpenAI API. If not present, the default value supplied from LangChain4j is used.
custom-headers	Map<string, string>	A map containing custom headers.
enabled	boolean	If set to false, the OpenAI model will not be available even if configured.
log-requests	boolean	Whether to log API requests.
log-responses	boolean	Whether to log API responses.
max-retries	integer	The maximum number of retries for failed API requests.
model-name	string	The model name to use.
organization-id	string	The ID of the organization for API requests.
temperature	double	The sampling temperature to use, between 0 and 2. Higher values make the output more random, while lower values make it more focused and deterministic.
timeout	duration	The timeout setting for API requests. See here for the format.

Proxy

The framework attempts to resolve a proxy using the following strategy:

1. Check for a named service: Look in the service registry for a service of type java.net.Proxy with the name open-ai.language-model.
2. Fallback to another named service: If the first service is not found, search for a java.net.Proxy service with the name open-ai.
3. No proxy: If neither service is found, the framework will not use a proxy.

Tokenizer

The framework attempts to resolve a tokenizer using the following strategy:

1. Check for a named service: Look in the service registry for a service of type dev.langchain4j.model.Tokenizer with the name open-ai.language-model.
2. Fallback to another named service: If the first service is not found, search for a dev.langchain4j.model.Tokenizer service with the name open-ai.
3. Default Tokenizer: If neither service is found, the framework will use the default tokenizer provided by LangChain4j.

OpenAiModerationModel

To automatically create and add OpenAiModerationModel to the service registry add the following lines to application.yaml:

langchain4j:
  providers:
    open-ai:
      api-key: "demo"

  models:
    openai-moderation-model:
      provider: open-ai
      model-name: "omni-moderation-latest"

If enabled is set to false, the configuration is ignored, and the component is not created.

Full list of configuration properties:

Key	Type	Description
api-key	string	Required. The API key used to authenticate requests to the OpenAI API.
base-url	string	The base URL for the OpenAI API. If not present, the default value supplied from LangChain4j is used.
custom-headers	Map<string, string>	A map containing custom headers.
enabled	boolean	If set to false, the OpenAI model will not be available even if configured.
log-requests	boolean	Whether to log API requests.
log-responses	boolean	Whether to log API responses.
max-retries	integer	The maximum number of retries for failed API requests.
model-name	string	The model name to use.
organization-id	string	The ID of the organization for API requests.
timeout	duration	The timeout setting for API requests. See here for the format.

Proxy

The framework attempts to resolve a proxy using the following strategy:

1. Check for a named service: Look in the service registry for a service of type java.net.Proxy with the name open-ai.moderation-model.
2. Fallback to another named service: If the first service is not found, search for a java.net.Proxy service with the name open-ai.
3. No proxy: If neither service is found, the framework will not use a proxy.

OpenAiStreamingChatModel

To automatically create and add OpenAiStreamingChatModel to the service registry add the following lines to application.yaml:

langchain4j:
  providers:
    open-ai:
      api-key: "demo"

  models:
    openai-streaming-chat-model:
      provider: open-ai
      model-name: "gpt-4o-mini"

If enabled is set to false, the configuration is ignored, and the component is not created.

Full list of configuration properties:

Key	Type	Description
api-key	string	Required. The API key used to authenticate requests to the OpenAI API.
base-url	string	The base URL for the OpenAI API. If not present, the default value supplied from LangChain4j is used.
custom-headers	Map<string, string>	A map containing custom headers.
enabled	boolean	If set to false, OpenAI model will not be available even if configured.
frequency-penalty	double	The frequency penalty, between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model’s likelihood to repeat the same line.
log-requests	boolean	Whether to log API requests.
log-responses	boolean	Whether to log API responses.
logit-bias	Map<string, int>	LogitBias adjusts the likelihood of specific tokens appearing in a model’s response. A map of token IDs to bias values (-100 to 100). Positive values increase the chance of the token, while negative values reduce it, allowing fine control over token preferences in the output.
max-completion-tokens	int	The maximum number of tokens allowed for the model’s response.
max-tokens	int	The maximum number of tokens to generate in the completion.
model-name	string	The model name to use (e.g., "gpt-3.5-turbo").
organization-id	string	The ID of the organization for API requests.
parallel-tool-calls	boolean	Whether to allow parallel calls to tools.
presence-penalty	double	The presence penalty, between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, encouraging the model to use new words.
response-format	string	The format in which the model should return the response.
seed	int	The seed for the random number generator used by the model.
stop	string[]	The list of sequences where the API will stop generating further tokens.
strict-tools	boolean	Whether to enforce strict validation of tools used by the model.
temperature	double	The sampling temperature to use, between 0 and 2. Higher values make the output more random, while lower values make it more focused and deterministic.
timeout	duration	The timeout setting for API requests. See here for the format.
top-p	double	The nucleus sampling value, where the model considers the results of the tokens with top_p probability mass.
user	string	The user ID associated with the API requests.

Proxy

The framework attempts to resolve a proxy using the following strategy:

1. Check for a named service: Look in the service registry for a service of type java.net.Proxy with the name open-ai.streaming-chat-model.
2. Fallback to another named service: If the first service is not found, search for a java.net.Proxy service with the name open-ai.
3. No proxy: If neither service is found, the framework will not use a proxy.

Tokenizer

The framework attempts to resolve a tokenizer using the following strategy:

1. Check for a named service: Look in the service registry for a service of type dev.langchain4j.model.Tokenizer with the name open-ai.streaming-chat-model.
2. Fallback to another named service: If the first service is not found, search for a dev.langchain4j.model.Tokenizer service with the name open-ai.
3. Default Tokenizer: If neither service is found, the framework will use the default tokenizer provided by LangChain4j.

Additional Information

• LangChain4j Integration
• LangChain4j OpenAI Documentation
• OpenAI API Documentation
• OpenAI API Reference