All Classes and Interfaces
Class
Description
Abstract base class for
GuardrailExecutors.A generic abstract builder class for creating instances of
GuardrailExecutor.Represents a message generated by AI (language model).
Builder for
Audio.Represents a request for
ChatMessage augmentation.Represents the result of a
ChatMessage augmentation.Represents a capability of a
ChatModel or StreamingChatModel.Generic executor interface that defines a chat interaction
An abstract base-builder class for constructing instances of
ChatExecutor.StreamingToSynchronousBuilder for constructing instances of
ChatExecutor.SynchronousBuilder for constructing instances of
ChatExecutor.Represents the memory (history) of a chat conversation.
Represents a store for the
ChatMemory state.Represents a chat message.
A deserializer for
ChatMessage objects.A codec for serializing and deserializing
ChatMessage objects to and from JSON.A factory for creating
ChatMessageJsonCodec objects.The type of content, e.g.
Represents a language model that has a chat API.
The chat model error context.
A
ChatModel listener that listens for requests, responses and errors.The chat model request context.
The chat response context.
Represents common chat request parameters supported by most LLM providers.
Represents common chat response metadata supported by most LLM providers.
A factory for providing instances of classes
Utility class for creating and retrieving instances of specified class types.
A factory interface for providing access to class metadata.
Interface for executing code.
Represents a complete tool call.
A
QueryTransformer that leverages a ChatModel to condense a given Query
along with a chat memory (previous conversation history) into a concise Query.A filter that checks if the value of a metadata key contains a specific string.
Abstract base interface for message content.
Represents content relevant to a user
Query with the potential to enhance and ground the LLM's response.Exception thrown when the LLM provider refuses to process a request due to content filtering
or violation of usage policies.
Injects given
Contents into a given UserMessage.The type of content, e.g.
Utility class for calculating cosine similarity between two vectors.
Represents a custom message.
Utility class to guess the mime-type of a file from its path or URI.
A default implementation of a
Content.Default implementation of
ContentAggregator intended to be suitable for the majority of use cases.Default implementation of
ContentInjector intended to be suitable for the majority of use cases.A default implementation of a
Document.Default implementation of
QueryRouter intended to be suitable for the majority of use cases.Default implementation of
QueryTransformer intended to be suitable for the majority of use cases.The default implementation of
RetrievalAugmentor intended to be suitable for the majority of use cases.Annotation to attach a description to a class field.
A dimension aware embedding model
A
ChatModel which throws a ModelDisabledException for all of its methodsAn
EmbeddingModel which throws a ModelDisabledException for all of its methodsAn
ImageModel which throws a ModelDisabledException for all of its methodsA
LanguageModel which throws a ModelDisabledException for all of its methodsA
ModerationModel which throws a ModelDisabledException for all of its methodsA
StreamingChatModel which throws a ModelDisabledException for all of its methodsA
StreamingLanguageModel which throws a ModelDisabledException for all of its methodsRepresents an unstructured piece of text that usually corresponds to a content of a single file.
Utility class for loading documents.
Defines the interface for parsing an
InputStream into a Document.A factory for creating
DocumentParser instances through SPI.Defines the interface for a Document source.
Defines the interface for splitting a document into text segments.
A factory for creating
DocumentSplitter instances through SPI.Defines the interface for transforming a
Document.Represents a dense vector embedding of a text.
Represents a matched embedding along with its relevance score (derivative of cosine distance), ID, and original embedded content.
Represents a model that can convert a given text into an embedding (vector representation of the text).
A factory for creating
EmbeddingModel instances through SPI.Represents a request to search in an
EmbeddingStore.Represents a result of a search in an
EmbeddingStore.Represents a store for embeddings, also known as a vector database.
A
ContentRetriever that retrieves from an EmbeddingStore.The
EmbeddingStoreIngestor represents an ingestion pipeline and is responsible
for ingesting Documents into an EmbeddingStore.EmbeddingStoreIngestor builder.
Utility methods for creating common exceptions.
Indicates that a class/constructor/method is experimental and might change in the future.
This class represents a filter that can be applied during search in an
EmbeddingStore.Parses a filter expression string into a
Filter object.The reason why a model call finished.
A guardrail is a rule that is applied when interacting with an LLM either to the input (the user message) or to the
output of the model to ensure that they are safe and meet the expectations of the model.
Exception thrown when an input or output guardrail validation fails.
Represents a mechanism to execute a set of guardrails on given parameters.
Represents a factory for creating instances of
AbstractGuardrailExecutor.GuardrailExecutorBuilder.Represents the parameter passed to
Guardrail.validate(GuardrailRequest)} in order to validate an interaction
between a user and the LLM.Represents the common parameters shared across guardrail checks when validating interactions
between a user and a language model.
Builder for
GuardrailRequestParams.The result of the validation of an interaction between a user and the LLM.
The message and the cause of the failure of a single validation.
The possible results of a guardrails validation.
Base interface for common configuration across all kinds of guardrails.
Builder for
GuardrailsConfig instances.Represents an image as a URL or as a Base64-encoded string.
Builder for
Image.Represents an image with a DetailLevel.
The detail level of an
Image.Text to Image generator model.
Represents the result of a
EmbeddingStoreIngestor ingestion process.Implementation of
ChatMemoryStore that stores state of ChatMemory (chat messages) in-memory.An input guardrail is a rule that is applied to the input of the model to ensure that the input (i.e.
Exception thrown when an input guardrail validation fails.
The
GuardrailExecutor for InputGuardrails.Builder class for constructing instances of
InputGuardrailExecutor.Represents a factory for creating instances of
InputGuardrailExecutor.InputGuardrailExecutorBuilder.Represents the parameter passed to
InputGuardrail.validate(InputGuardrailRequest).Builder for
InputGuardrailRequest.The result of the validation of an
InputGuardrailRepresents an input guardrail failure
Configuration specifically for input guardrails.
Builder for
InputGuardrailsConfig instances.SPI for overriding and/or extending the default
InputGuardrailsConfig.InputGuardrailsConfigBuilder implementation.Indicates that the annotated class is intended for internal use only within the library.
Annotation to mark methods where JaCoCo coverage should be ignored.
JSON helper class.
The abstract JSON codec interface.
A factory for creating
Json.JsonCodec instances through SPI.An
OutputGuardrail that will check whether or not a response can be successfully deserialized to an object
of type T from JSONCan reference
JsonObjectSchema when recursion is required.A base interface for a JSON schema element.
Utility class with lambda-based streaming response handlers.
Represents a language model that has a simple text interface (as opposed to a chat interface).
A
QueryRouter that utilizes a ChatModel to make a routing decision.Strategy applied if the call to the LLM fails of if LLM does not return a valid response.
Represents metadata of a
Document or a TextSegment.Represents metadata that may be useful or necessary for retrieval or augmentation purposes.
An exception thrown by a model that could be disabled by a user.
Represents moderation status.
Represents a model that can moderate text.
An output guardrail is a rule that is applied to the output of the model to ensure that the output is safe and meets
the expectations.
Exception thrown when an output guardrail validation fails.
The
GuardrailExecutor for OutputGuardrails.Builder class for constructing instances of
OutputGuardrailExecutor.Represents a factory for creating instances of
OutputGuardrailExecutor.OutputGuardrailExecutorBuilder.Represents the parameter passed to
OutputGuardrail.validate(OutputGuardrailRequest).Builder for
OutputGuardrailRequest.The result of the validation of an
OutputGuardrailRepresents an output guardrail failure
Configuration specifically for output guardrails.
Builder for
OutputGuardrailsConfig instances.SPI for overriding and/or extending the default
OutputGuardrailsConfig.OutputGuardrailsConfigBuilder implementation.Parameter of a Tool
Represents a partial tool call.
Builder for
PdfFile.Represents a prompt (an input text sent to the LLM).
Represents a template of a prompt that can be reused multiple times.
A factory for creating prompt templates.
Interface for input for the factory.
Interface for a prompt template.
Represents a query from the user intended for retrieving relevant
Contents.Routes the given
Query to one or multiple ContentRetrievers.Implementation of Reciprocal Rank Fusion.
Utility class for converting between cosine similarity and relevance score.
A
ContentAggregator that performs re-ranking using a ScoringModel, such as Cohere.Represents the response from various types of models, including language, chat, embedding, and moderation models.
Augments the provided
ChatMessage with retrieved Contents.Utility class for retrying actions.
This class encapsulates a retry policy.
This class encapsulates a retry policy builder.
Represents a model capable of scoring a text against a query.
Utility wrapper around
ServiceLoader.load().Represents a language model that has a chat API and can stream a response one token at a time.
Represents a handler for a
StreamingChatModel response.Represents a language model that has a simple text interface (as opposed to a chat interface)
and can stream a response one token at a time.
Represents a handler for streaming responses from a language model.
Represents a structured prompt.
Utility class for
StructuredPrompt.Represents a factory for structured prompts.
Utility class for structured prompts.
Represents a system message, typically defined by a developer.
Represents a text content.
Represents a semantically meaningful segment (chunk/piece/fragment) of a larger entity such as a document or chat conversation.
Defines the interface for transforming a
TextSegment.Represents an interface for estimating the count of tokens in various text types such as a text, prompt, text segment, etc.
Represents the token usage of a response.
Java methods annotated with
@Tool are considered tools/functions that language model can execute/call.Internal helper that helps to build partial and complete tool calls during streaming.
Specifies how
ChatModel should use tools.Represents an LLM-generated request to execute a tool.
ToolExecutionRequest builder static inner class.Represents the result of a tool execution in response to a
ToolExecutionRequest.If a
Tool method parameter is annotated with this annotation,
memory id (parameter annotated with @MemoryId in AI Service) will be injected automatically.Describes a tool that language model can execute.
ToolSpecification builder static inner class.Utility methods for
ToolSpecifications.Represents a message from a user, typically an end user of the application.
Utility methods.
Utility class for validating method arguments.
Builder for
Video.Utility class for working with virtual threads introduced in Java 21.
Represents a web search engine that can be used to perform searches on the Web in response to a user query.
Represents general information about the web search performed.
Represents an organic search results are the web pages that are returned by the search engine in response to a search query.
Represents a search request that can be made by the user to perform searches in any implementation of
WebSearchEngine.Represents the response of a web search performed.