Uses of Annotation Interface dev.langchain4j.Experimental (LangChain4j

Uses of Experimental in dev.langchain4j.agent.tool

Classes in dev.langchain4j.agent.tool with annotations of type Experimental

Modifier and Type

Class

Description

@interface

CompensateFor

Marks a method as the compensating action for a Tool.

enum

ReturnBehavior

Per-tool setting controlling what happens with a tool's result after execution.

enum

SearchBehavior

Defines the behavior of a tool when dev.langchain4j.service.tool.search.ToolSearchStrategy is configured for an AI Service.
Uses of Experimental in dev.langchain4j.agentic.patterns.cascade

Classes in dev.langchain4j.agentic.patterns.cascade with annotations of type Experimental

Modifier and Type

Class

Description

class

ConfidenceCascadePlanner

A planner trying its subagents one at a time, from the cheapest to the most capable, until one of them gives an answer that a DecisionModel accepts.
Uses of Experimental in dev.langchain4j.agentic.patterns.decisionrouter

Classes in dev.langchain4j.agentic.patterns.decisionrouter with annotations of type Experimental

Modifier and Type

Class

Description

class

DecisionRouterPlanner

A router planner that lets a DecisionModel choose which subagents handle a request.
Uses of Experimental in dev.langchain4j.data.message

Methods in dev.langchain4j.data.message with annotations of type Experimental

Modifier and Type

Method

Description

<T> T

AiMessage.attribute(String key, Class<T> type)

Returns additional attribute by it's key.

<T> T

SystemMessage.attribute(String key, Class<T> type)

Returns additional attribute by it's key.

<T> T

UserMessage.attribute(String key, Class<T> type)

Returns additional attribute by it's key.

Map<String,Object>

AiMessage.attributes()

Returns additional attributes, typically provider-specific.

AiMessage.Builder

AiMessage.Builder.attributes(Map<String,Object> attributes)

Map<String,Object>

SystemMessage.attributes()

Returns additional attributes.

Map<String,Object>

UserMessage.attributes()

Returns additional attributes.

List<Audio>

AiMessage.audios()

Get the audio generated by the model, for example by a text-to-speech model.

ToolExecutionResultMessage.Builder

ToolExecutionResultMessage.Builder.contents(Content... contents)

Sets the contents of the tool execution result.

ToolExecutionResultMessage.Builder

ToolExecutionResultMessage.Builder.contents(List<Content> contents)

Sets the contents of the tool execution result.

List<Content>

ToolExecutionResultMessage.contents()

Returns the Contents of the tool execution result.

boolean

ToolExecutionResultMessage.hasSingleText()

Whether this message contains a single TextContent.

AiMessage.Builder

AiMessage.Builder.thinking(String thinking)

String

AiMessage.thinking()

Get the thinking/reasoning text of the message.
Uses of Experimental in dev.langchain4j.exception

Classes in dev.langchain4j.exception with annotations of type Experimental

Modifier and Type

Class

Description

class

AsyncNotSupportedException

Signals that an asynchronous SPI method (an *Async method, or a reactive Publisher-returning method) is not implemented by a component, and is therefore the internal "this component is not genuinely asynchronous" marker carried by those methods' default implementations.

class

InvalidDecisionResponseException

Thrown when a DecisionModel returns a response that does not match the request, for example when an answer is missing, has the wrong type, or chooses an option that was not offered.
Uses of Experimental in dev.langchain4j.experimental.rag.content.retriever.sql

Classes in dev.langchain4j.experimental.rag.content.retriever.sql with annotations of type Experimental

Modifier and Type

Class

Description

class

SqlDatabaseContentRetriever

WARNING!

Constructors in dev.langchain4j.experimental.rag.content.retriever.sql with annotations of type Experimental

Modifier

Constructor

Description

SqlDatabaseContentRetriever(DataSource dataSource, String sqlDialect, String databaseStructure, PromptTemplate promptTemplate, ChatModel chatModel, Integer maxRetries)

Creates an instance of a SqlDatabaseContentRetriever.
Uses of Experimental in dev.langchain4j.guardrail

Methods in dev.langchain4j.guardrail with annotations of type Experimental

Modifier and Type

Method

Description

OutputGuardrailRequest.Builder

OutputGuardrailRequest.Builder.chatRequest(ChatRequest chatRequest)

Sets the request that was sent to the model to produce the response.

ChatRequest

OutputGuardrailRequest.chatRequest()

Returns the request that was sent to the model to produce OutputGuardrailRequest.responseFromLLM(), if known: when the response is the result of a tool-calling loop, the last request of the loop, whose messages include the tool calls and tool results of all rounds; when the model was asked again after a guardrail failure (reprompt), the request of that call, whose messages include the reprompt.

default ChatExchange

ChatExecutor.exchange(List<ChatMessage> chatMessages)

Like ChatExecutor.execute(List), but also returns the request that produced the response.

default CompletableFuture<ChatExchange>

ChatExecutor.exchangeAsync(List<ChatMessage> chatMessages)

Non-blocking counterpart of ChatExecutor.exchange(List).

default CompletableFuture<ChatResponse>

ChatExecutor.executeAsync()

Non-blocking counterpart of ChatExecutor.execute().

default CompletableFuture<ChatResponse>

ChatExecutor.executeAsync(List<ChatMessage> chatMessages)

Non-blocking counterpart of ChatExecutor.execute(List).

default CompletableFuture<R>

Guardrail.validateAsync(P request)

Non-blocking counterpart of Guardrail.validate(GuardrailRequest), invoked by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service modes.
Uses of Experimental in dev.langchain4j.guardrails

Classes in dev.langchain4j.guardrails with annotations of type Experimental

Modifier and Type

Class

Description

class

DecisionModelInputGuardrail

An InputGuardrail that checks user messages with a DecisionModel.

class

DecisionModelOutputGuardrail

An OutputGuardrail that checks the responses of the model with a DecisionModel.
Uses of Experimental in dev.langchain4j.http.client

Classes in dev.langchain4j.http.client with annotations of type Experimental

Modifier and Type

Class

Description

class

FormDataFile

Methods in dev.langchain4j.http.client with annotations of type Experimental

Modifier and Type

Method

Description

HttpRequest.Builder

HttpRequest.Builder.addFormDataField(String name, String value)

HttpRequest.Builder

HttpRequest.Builder.addFormDataFile(String name, String fileName, String contentType, byte[] content)

default CompletableFuture<SuccessfulHttpResponse>

HttpClient.executeAsync(HttpRequest request)

Non-blocking counterpart of HttpClient.execute(HttpRequest).

HttpRequest.Builder

HttpRequest.Builder.formDataFields(Map<String,String> formDataFields)

Map<String,String>

HttpRequest.formDataFields()

HttpRequest.Builder

HttpRequest.Builder.formDataFiles(Map<String, FormDataFile> formDataFiles)

Map<String, FormDataFile>

HttpRequest.formDataFiles()

default Flow.Publisher<HttpStreamingEvent>

HttpClient.stream(HttpRequest request)

Executes a streaming HTTP request and exposes the parsed events as a cold Flow.Publisher of HttpStreamingEvents.

default Flow.Publisher<HttpStreamingEvent>

HttpClient.stream(HttpRequest request, ServerSentEventParser parser)

Like HttpClient.stream(HttpRequest), but with a caller-supplied ServerSentEventParser.
Uses of Experimental in dev.langchain4j.http.client.apache

Methods in dev.langchain4j.http.client.apache with annotations of type Experimental

Modifier and Type

Method

Description

Integer

ApacheHttpClientBuilder.streamingBufferSize()

ApacheHttpClientBuilder

ApacheHttpClientBuilder.streamingBufferSize(Integer streamingBufferSize)

Sets the size of the bounded back-pressure buffer used by the streaming (stream) path.
Uses of Experimental in dev.langchain4j.http.client.jdk

Methods in dev.langchain4j.http.client.jdk with annotations of type Experimental

Modifier and Type

Method

Description

Integer

JdkHttpClientBuilder.streamingBufferSize()

JdkHttpClientBuilder

JdkHttpClientBuilder.streamingBufferSize(Integer streamingBufferSize)

Sets the size of the bounded back-pressure buffer used by the streaming (stream) path.
Uses of Experimental in dev.langchain4j.http.client.okhttp

Methods in dev.langchain4j.http.client.okhttp with annotations of type Experimental

Modifier and Type

Method

Description

Integer

OkHttpClientBuilder.streamingBufferSize()

OkHttpClientBuilder

OkHttpClientBuilder.streamingBufferSize(Integer streamingBufferSize)

Sets the size of the bounded back-pressure buffer used by the streaming (stream) path.
Uses of Experimental in dev.langchain4j.http.client.sse

Classes in dev.langchain4j.http.client.sse with annotations of type Experimental

Modifier and Type

Class

Description

class

HttpResponseReceived

HttpStreamingEvent signalling that the HTTP response head (status and headers) was received and the stream opened successfully.

interface

HttpStreamingEvent

An event emitted while consuming a streaming HTTP response via HttpClient.stream(dev.langchain4j.http.client.HttpRequest).

class

ServerSentEventContext

interface

ServerSentEventParsingHandle

Handle that can be used to cancel the parsing of server-sent events.

Methods in dev.langchain4j.http.client.sse with annotations of type Experimental

Modifier and Type

Method

Description

default ServerSentEventParser.Incremental

ServerSentEventParser.incremental()

Returns a fresh ServerSentEventParser.Incremental parser instance backed by the same parsing logic as ServerSentEventParser.parse(InputStream, ServerSentEventListener), but driven by byte chunks instead of a blocking InputStream.

default void

ServerSentEventListener.onEvent(ServerSentEvent event, ServerSentEventContext context)

Handles server-sent event.
Uses of Experimental in dev.langchain4j.invocation

Classes in dev.langchain4j.invocation with annotations of type Experimental

Modifier and Type

Class

Description

final class

InvocationEnd

Makes sure that an AI Service invocation reports its end only once, with exactly one of AiServiceCompletedEvent, AiServiceErrorEvent or AiServiceCancelledEvent, also when completion, failure and cancellation race.
Uses of Experimental in dev.langchain4j.mcp.client

Methods in dev.langchain4j.mcp.client with annotations of type Experimental

Modifier and Type

Method

Description

default CompletableFuture<ToolExecutionResult>

McpClient.executeToolAsync(ToolExecutionRequest executionRequest, InvocationContext invocationContext)

Non-blocking counterpart of McpClient.executeTool(ToolExecutionRequest, InvocationContext): executes a tool on the MCP server without holding a thread while the result is in flight.
Uses of Experimental in dev.langchain4j.memory

Methods in dev.langchain4j.memory with annotations of type Experimental

Modifier and Type

Method

Description

default CompletableFuture<Void>

ChatMemory.addAsync(List<ChatMessage> messages)

Non-blocking counterpart of ChatMemory.add(Iterable), used by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service APIs.

default CompletableFuture<List<ChatMessage>>

ChatMemory.messagesAsync()

Non-blocking counterpart of ChatMemory.messages(), used by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service APIs.

default CompletableFuture<Void>

ChatMemory.setAsync(List<ChatMessage> messages)

Non-blocking counterpart of ChatMemory.set(Iterable), used by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service APIs (e.g. to rewrite memory for tool compensation without blocking the model-delivery thread).
Uses of Experimental in dev.langchain4j.micrometer.metrics.listeners

Classes in dev.langchain4j.micrometer.metrics.listeners with annotations of type Experimental

Modifier and Type

Class

Description

class

MicrometerMetricsChatModelListener

A ChatModelListener that uses a Micrometer MeterRegistry to collect metrics about chat model interactions following OpenTelemetry Semantic Conventions for Generative AI.

class

MicrometerMetricsEmbeddingModelListener

An EmbeddingModelListener that uses a Micrometer MeterRegistry to collect metrics about embedding model interactions following OpenTelemetry Semantic Conventions for Generative AI.
Uses of Experimental in dev.langchain4j.model.anthropic

Classes in dev.langchain4j.model.anthropic with annotations of type Experimental

Modifier and Type

Class

Description

final class

AnthropicBatchChatModel

A BatchChatModel for the Anthropic Message Batches API, which processes multiple chat requests asynchronously at 50% of the standard per-token price.

class

AnthropicCacheDiagnostics

Result of Anthropic's (beta) cache diagnostics comparison for a single request, surfaced via AnthropicChatResponseMetadata.cacheDiagnostics().

class

AnthropicServerTool

class

AnthropicServerToolResult

Represents a result from an Anthropic server-executed tool (e.g., web_search, code_execution).

enum

AnthropicSkill

Anthropic-managed Agent Skills that let Claude generate real downloadable documents by running pre-built skills inside the code execution container.

class

AnthropicTokenCountEstimator

Methods in dev.langchain4j.model.anthropic with annotations of type Experimental

Modifier and Type

Method

Description

AnthropicStreamingChatModel.AnthropicStreamingChatModelBuilder

AnthropicStreamingChatModel.AnthropicStreamingChatModelBuilder.streamingBufferSize(Integer streamingBufferSize)

Sets the size of the bounded back-pressure buffer for the reactive (Flow.Publisher) streaming path.
Uses of Experimental in dev.langchain4j.model.audio

Classes in dev.langchain4j.model.audio with annotations of type Experimental

Modifier and Type

Interface

Description

interface

AudioTranscriptionModel

A model that can transcribe audio into text.

class

AudioTranscriptionRequest

Request to transcribe audio.

class

AudioTranscriptionResponse

Response containing the transcription of an audio file.

interface

TextToSpeechModel

A model that can synthesize audio from text.

class

TextToSpeechRequest

class

TextToSpeechResponse

Response containing the audio generated from text.
Uses of Experimental in dev.langchain4j.model.azure

Classes in dev.langchain4j.model.azure with annotations of type Experimental

Modifier and Type

Class

Description

class

AzureOpenAiAudioTranscriptionModel

Represents an Azure OpenAI audio transcription model, such as Whisper.
Uses of Experimental in dev.langchain4j.model.batch

Classes in dev.langchain4j.model.batch with annotations of type Experimental

Modifier and Type

Class

Description

final class

BatchError

Represents an error status for a failed request within a batch operation.

interface

BatchItemResult<T>

Represents the outcome of a single request within a batch operation.

class

BatchPage<T>

Represents a set of batch jobs that is potentially paginated.

final class

BatchPagination

class

BatchRequest<T>

Represents a batch of requests to be processed together.

class

BatchResponse<T>

Represents the responses of a batch operation.

enum

BatchState

Represents the possible states of a batch job.
Uses of Experimental in dev.langchain4j.model.bedrock

Classes in dev.langchain4j.model.bedrock with annotations of type Experimental

Modifier and Type

Class

Description

final class

BedrockBatchChatModel

A BatchChatModel for the Amazon Bedrock batch inference API, which processes multiple chat requests asynchronously at a lower price than on-demand inference for supported models.

Methods in dev.langchain4j.model.bedrock with annotations of type Experimental

Modifier and Type

Method

Description

BedrockChatModel.Builder

BedrockChatModel.Builder.asyncClient(software.amazon.awssdk.services.bedrockruntime.BedrockRuntimeAsyncClient asyncClient)

Sets the async Bedrock client used by chatAsync.

BedrockStreamingChatModel.Builder

BedrockStreamingChatModel.Builder.streamingBufferSize(Integer streamingBufferSize)

Sets the size of the bounded back-pressure buffer for the reactive (Flow.Publisher) streaming path.
Uses of Experimental in dev.langchain4j.model.catalog

Classes in dev.langchain4j.model.catalog with annotations of type Experimental

Modifier and Type

Interface

Description

interface

ModelCatalog

Represents a service that can discover available models from an LLM provider.

class

ModelDescription

Represents metadata about an available model from a provider.

enum

ModelType

Represents the type/category of a model.
Uses of Experimental in dev.langchain4j.model.chat

Classes in dev.langchain4j.model.chat with annotations of type Experimental

Modifier and Type

Interface

Description

interface

BatchChatModel

A model interface for processing multiple chat requests asynchronously in a batch.

final class

ChatExchange

A request sent to a chat model and the response it returned.

Methods in dev.langchain4j.model.chat with annotations of type Experimental

Modifier and Type

Method

Description

default Flow.Publisher<ChatModelStreamingEvent>

StreamingChatModel.chat(ChatMessage... messages)

Reactive convenience overload accepting the messages directly: returns a cold Publisher that streams the response to messages.

default Flow.Publisher<ChatModelStreamingEvent>

StreamingChatModel.chat(ChatRequest request)

Reactive entry point: sends a chat request and returns a Flow.Publisher of ChatModelStreamingEvents.

default Flow.Publisher<String>

StreamingChatModel.chat(String userMessage)

Reactive convenience counterpart of StreamingChatModel.chat(String, StreamingChatResponseHandler): returns a cold Publisher that streams the model's textual response to userMessage, token by token.

default Flow.Publisher<ChatModelStreamingEvent>

StreamingChatModel.chat(List<ChatMessage> messages)

Reactive convenience counterpart of StreamingChatModel.chat(List, StreamingChatResponseHandler): returns a cold Publisher that streams the response to messages.

default CompletableFuture<ChatResponse>

ChatModel.chatAsync(ChatMessage... messages)

Non-blocking convenience counterpart of ChatModel.chat(ChatMessage...).

default CompletableFuture<ChatResponse>

ChatModel.chatAsync(ChatRequest chatRequest)

Non-blocking counterpart of ChatModel.chat(ChatRequest): sends a chat request and returns a CompletableFuture that completes with the ChatResponse once the model responds.

default CompletableFuture<ChatResponse>

ChatModel.chatAsync(ChatRequest chatRequest, ChatRequestOptions options)

Sends a non-blocking chat request with additional invocation options.

default CompletableFuture<String>

ChatModel.chatAsync(String userMessage)

Non-blocking convenience counterpart of ChatModel.chat(String): completes with the text of the model's response.

default CompletableFuture<ChatResponse>

ChatModel.chatAsync(List<ChatMessage> messages)

Non-blocking convenience counterpart of ChatModel.chat(List).

default Flow.Publisher<ChatModelStreamingEvent>

StreamingChatModel.doChat(ChatRequest chatRequest)

Provider-specific implementation of the reactive stream returned by StreamingChatModel.chat(ChatRequest) (which wraps it with ChatModelListener invocation).

default CompletableFuture<ChatResponse>

ChatModel.doChatAsync(ChatRequest chatRequest)

SPI hook for a genuinely non-blocking chat implementation, invoked by ChatModel.chatAsync(ChatRequest).
Uses of Experimental in dev.langchain4j.model.chat.response

Classes in dev.langchain4j.model.chat.response with annotations of type Experimental

Modifier and Type

Interface

Description

interface

ChatModelStreamingEvent

Base type for the events emitted by the reactive stream of a StreamingChatModel — the Flow.Publisher<ChatModelStreamingEvent> returned by StreamingChatModel.chat(dev.langchain4j.model.chat.request.ChatRequest).

class

CompleteResponse

The terminal ChatModelStreamingEvent of a streaming chat: the complete, aggregated ChatResponse.

class

CompleteToolCall

Represents a complete tool call.

class

PartialResponse

class

PartialResponseContext

class

PartialThinking

class

PartialThinkingContext

class

PartialToolCall

Represents a partial tool call.

class

PartialToolCallContext

interface

RawStreamingEvent

A provider-specific streaming event that langchain4j's generic event model does not (yet) map to a dedicated ChatModelStreamingEvent type — for example a web-search, citation or reasoning-summary event.

interface

StreamingHandle

Handle that can be used to cancel the streaming done via StreamingChatResponseHandler.

Methods in dev.langchain4j.model.chat.response with annotations of type Experimental

Modifier and Type

Method

Description

default void

StreamingChatResponseHandler.onCompleteToolCall(CompleteToolCall completeToolCall)

Invoked when the model has finished streaming a single tool call.

default void

StreamingChatResponseHandler.onPartialResponse(PartialResponse partialResponse, PartialResponseContext context)

Invoked each time the model generates a partial textual response, usually a single token.

default void

StreamingChatResponseHandler.onPartialThinking(PartialThinking partialThinking)

Invoked each time the model generates a partial thinking/reasoning text, usually a single token.

default void

StreamingChatResponseHandler.onPartialThinking(PartialThinking partialThinking, PartialThinkingContext context)

Invoked each time the model generates a partial thinking/reasoning text, usually a single token.

default void

StreamingChatResponseHandler.onPartialToolCall(PartialToolCall partialToolCall)

This callback is invoked each time the model generates a partial tool call, which contains a single token of the tool's arguments.

default void

StreamingChatResponseHandler.onPartialToolCall(PartialToolCall partialToolCall, PartialToolCallContext context)

This callback is invoked each time the model generates a partial tool call, which contains a single token of the tool's arguments.

default void

StreamingChatResponseHandler.onUnmappedRawEvent(Object rawEvent)

Invoked when a provider emits a raw streaming event that is not already exposed through one of the typed callbacks (such as StreamingChatResponseHandler.onPartialResponse(PartialResponse, PartialResponseContext), StreamingChatResponseHandler.onPartialThinking(PartialThinking, PartialThinkingContext), StreamingChatResponseHandler.onPartialToolCall(PartialToolCall, PartialToolCallContext) or StreamingChatResponseHandler.onCompleteToolCall(CompleteToolCall)).
Uses of Experimental in dev.langchain4j.model.chat.router

Classes in dev.langchain4j.model.chat.router with annotations of type Experimental

Modifier and Type

Class

Description

final class

ChatModelRoute

A route of a RoutingChatModel or RoutingStreamingChatModel, as seen by a ChatModelRouter.

interface

ChatModelRouter

Decides which route of a RoutingChatModel or RoutingStreamingChatModel handles a chat request.

final class

ChatModelRoutingRequest

The input of a ChatModelRouter: the chat request to route, the routes that can handle it and the options of the call.

final class

ChatModelRoutingResult

The result of a ChatModelRouter: the route that handles the request, or the default route of the routing chat model.

class

DecisionModelChatModelRouter

A ChatModelRouter that uses a DecisionModel to select the route: the decision model chooses the route whose description fits the last user message best.

class

RoutingChatModel

A ChatModel that sends each request to one of several chat models, as decided by a ChatModelRouter.

class

RoutingStreamingChatModel

The streaming counterpart of RoutingChatModel: a StreamingChatModel that sends each request to one of several streaming chat models, as decided by a ChatModelRouter.
Uses of Experimental in dev.langchain4j.model.decision

Classes in dev.langchain4j.model.decision with annotations of type Experimental

Modifier and Type

Interface

Description

interface

DecisionModel

A model that evaluates an input against a set of named, typed questions and returns one typed answer per question, instead of generated text.
Uses of Experimental in dev.langchain4j.model.decision.listener

Classes in dev.langchain4j.model.decision.listener with annotations of type Experimental

Modifier and Type

Class

Description

class

DecisionModelErrorContext

The context of an error of a decision model: the error, the corresponding request, the model provider and attributes.

interface

DecisionModelListener

A DecisionModel listener that listens for requests, responses and errors, for example to log decisions for audit, or to record metrics and traces.

class

DecisionModelRequestContext

The context of a request to a decision model: the request, the model provider and attributes.

class

DecisionModelResponseContext

The context of a response from a decision model: the response, the corresponding request, the model provider and attributes.
Uses of Experimental in dev.langchain4j.model.decision.request

Classes in dev.langchain4j.model.decision.request with annotations of type Experimental

Modifier and Type

Class

Description

final class

ChoiceQuestion

A question that selects exactly one option out of a named set, answered with the chosen option and the probability of each option (see ChoiceAnswer).

final class

ContentsInput

A DecisionInput that is a list of contents, for example a TextContent and an ImageContent.

interface

DecisionInput

The input of a DecisionRequest: what the questions are answered about.

enum

DecisionInputType

The kind of a DecisionInput.

final class

DecisionRequest

A request to a DecisionModel: the DecisionRequest.input() to evaluate and the named DecisionRequest.questions() to answer about it.

interface

DecisionRequestParameters

The per-call parameters of a DecisionRequest, such as the DecisionRequestParameters.modelName() to use.

class

DefaultDecisionRequestParameters

The default implementation of DecisionRequestParameters.

final class

FieldsInput

A DecisionInput made of named fields, for example a support ticket together with the customer's plan: Map.of("ticket", ticket, "plan", "enterprise").

final class

MessagesInput

A DecisionInput that is a conversation: messages in the order they were exchanged, for example a chat between a user and an AI assistant, including the tools the assistant called (AiMessage.toolExecutionRequests()) and their results (ToolExecutionResultMessage).

interface

Question

A question asked to a DecisionModel.

final class

ScaleQuestion

A question that places the input on an ordered scale, answered with the mean level and the probability of each level (see ScaleAnswer).

final class

TextInput

A DecisionInput that is plain text, for example a support ticket.

final class

YesNoQuestion

A yes/no question, answered with the probability that the answer is "yes" (see YesNoAnswer).
Uses of Experimental in dev.langchain4j.model.decision.response

Classes in dev.langchain4j.model.decision.response with annotations of type Experimental

Modifier and Type

Class

Description

final class

ChoiceAnswer

The answer to a ChoiceQuestion: the chosen option, the probability of each option and, if the model reports one, a confidence.

interface

DecisionAnswer

An answer to one question of a DecisionRequest.

final class

DecisionResponse

The response of a DecisionModel: one DecisionAnswer per question of the request, keyed by question name, together with the DecisionResponseMetadata (model name, token usage).

class

DecisionResponseMetadata

The metadata of a DecisionResponse: the DecisionResponseMetadata.modelName() that produced the answers and the DecisionResponseMetadata.tokenUsage() the call consumed.

final class

RefusalAnswer

The answer to a question that the decision model refused to answer, for example because the input or the question goes against the usage policies of the provider.

final class

ScaleAnswer

The answer to a ScaleQuestion: the mean level, the probability of each level and, if the model reports one, a confidence.

final class

YesNoAnswer

The answer to a YesNoQuestion: the YesNoAnswer.probability() that the answer is "yes".
Uses of Experimental in dev.langchain4j.model.embedding

Classes in dev.langchain4j.model.embedding with annotations of type Experimental

Modifier and Type

Interface

Description

interface

BatchEmbeddingModel

Used for processing multiple embedding requests asynchronously in a batch.

Methods in dev.langchain4j.model.embedding with annotations of type Experimental

Modifier and Type

Method

Description

default EmbeddingModel

EmbeddingModel.addListener(EmbeddingModelListener listener)

Returns an EmbeddingModel that wraps this one and notifies the given listener around each embedding call.

default EmbeddingModel

EmbeddingModel.addListeners(List<EmbeddingModelListener> listeners)

Returns an EmbeddingModel that wraps this one and notifies the given listeners (in iteration order) around each embedding call.

default EmbeddingRequestParameters

EmbeddingModel.defaultRequestParameters()

The parameters applied to every request unless overridden by the request itself, typically derived from the model's builder-time configuration.

default EmbeddingResponse

EmbeddingModel.doEmbed(EmbeddingRequest request)

Performs the embedding for EmbeddingModel.embed(EmbeddingRequest).

default CompletableFuture<EmbeddingResponse>

EmbeddingModel.doEmbedAsync(EmbeddingRequest request)

Non-blocking counterpart of EmbeddingModel.doEmbed(EmbeddingRequest), called by EmbeddingModel.embedAsync(EmbeddingRequest) after the request's parameters have been merged and validated.

default EmbeddingResponse

EmbeddingModel.embed(EmbeddingRequest request)

Embeds the inputs of the given EmbeddingRequest and returns the resulting embeddings.

default CompletableFuture<EmbeddingResponse>

EmbeddingModel.embedAsync(EmbeddingRequest request)

Non-blocking counterpart of EmbeddingModel.embed(EmbeddingRequest): embeds the request's inputs and completes the returned future with the embeddings, without blocking the calling thread.

default List<EmbeddingModelListener>

EmbeddingModel.listeners()

The EmbeddingModelListeners that are notified around EmbeddingModel.embed(EmbeddingRequest).

default ModelProvider

EmbeddingModel.provider()

The ModelProvider of this embedding model (for example ModelProvider.OPEN_AI).

default Set<ContentType>

EmbeddingModel.supportedContentTypes()

The input content types this model can embed.

default Set<EmbeddingParameter<?>>

EmbeddingModel.supportedParameters()

The per-call parameters this model honors.
Uses of Experimental in dev.langchain4j.model.embedding.listener

Classes in dev.langchain4j.model.embedding.listener with annotations of type Experimental

Modifier and Type

Class

Description

class

EmbeddingModelErrorContext

The context passed to EmbeddingModelListener.onError(EmbeddingModelErrorContext) when an embedding call fails.

static class

EmbeddingModelErrorContext.Builder

Builder for EmbeddingModelErrorContext.

interface

EmbeddingModelListener

An EmbeddingModel listener that listens for requests, responses and errors.

class

EmbeddingModelRequestContext

The context passed to EmbeddingModelListener.onRequest(EmbeddingModelRequestContext) before an embedding call.

static class

EmbeddingModelRequestContext.Builder

Builder for EmbeddingModelRequestContext.

class

EmbeddingModelResponseContext

The context passed to EmbeddingModelListener.onResponse(EmbeddingModelResponseContext) after a successful embedding call.

static class

EmbeddingModelResponseContext.Builder

Builder for EmbeddingModelResponseContext.
Uses of Experimental in dev.langchain4j.model.embedding.request

Classes in dev.langchain4j.model.embedding.request with annotations of type Experimental

Modifier and Type

Class

Description

class

DefaultEmbeddingRequestParameters

The default, map-backed implementation of EmbeddingRequestParameters.

class

EmbeddingInput

A single input to embed: one or more ordered Content parts that are fused into one Embedding.

enum

EmbeddingInputType

The role a piece of text plays in a retrieval task, allowing a provider to encode it differently.

final record

EmbeddingParameter<T>

A typed, stable identifier for a single EmbeddingRequestParameters value.

class

EmbeddingRequest

A request to an EmbeddingModel, containing a batch of EmbeddingInputs to embed and the per-call parameters.

interface

EmbeddingRequestParameters

The per-call parameters of an EmbeddingRequest, such as EmbeddingRequestParameters.dimensions() or EmbeddingRequestParameters.inputType().
Uses of Experimental in dev.langchain4j.model.embedding.response

Classes in dev.langchain4j.model.embedding.response with annotations of type Experimental

Modifier and Type

Class

Description

class

EmbeddingResponse

The result of embedding an EmbeddingRequest: the Embeddings produced for the request's inputs (in the same order) together with the EmbeddingResponseMetadata such as the model name and token usage.

class

EmbeddingResponseMetadata

Metadata returned alongside an EmbeddingResponse, describing the call that produced it: the model name and the token usage.
Uses of Experimental in dev.langchain4j.model.google.genai

Classes in dev.langchain4j.model.google.genai with annotations of type Experimental

Modifier and Type

Class

Description

final class

GoogleGenAiBatchChatModel

Provides an interface for interacting with the Google GenAI Batch API for Chat models.

final class

GoogleGenAiBatchEmbeddingModel

Provides an interface for interacting with the Google GenAI Batch API for Embedding models.

final class

GoogleGenAiBatchImageModel

Provides an interface for interacting with the Google GenAI Batch API for Image generation models.

class

GoogleGenAiChatModel

class

GoogleGenAiEmbeddingModel

class

GoogleGenAiImageModel

Represents a Google GenAI model for image generation and editing using the official com.google.genai SDK.

class

GoogleGenAiStreamingChatModel

class

GoogleGenAiTextToSpeechModel

Represents a Google GenAI text-to-speech model, such as gemini-3.8-flash-tts, using the official com.google.genai SDK.
Uses of Experimental in dev.langchain4j.model.googleai

Classes in dev.langchain4j.model.googleai with annotations of type Experimental

Modifier and Type

Class

Description

final class

GoogleAiGeminiBatchChatModel

Provides an interface for interacting with the Gemini Batch API, an asynchronous service designed for processing large volumes of requests at a reduced cost (50% of standard).

final class

GoogleAiGeminiBatchEmbeddingModel

Batch embedding model for Google AI Gemini.

final class

GoogleAiGeminiBatchImageModel

Provides an interface for batch image generation using the Gemini Batch API.

class

GoogleAiGeminiImageModel

Represents a Google AI Gemini model for image generation and editing.
Uses of Experimental in dev.langchain4j.model.image

Classes in dev.langchain4j.model.image with annotations of type Experimental

Modifier and Type

Interface

Description

interface

BatchImageModel

A model interface for processing multiple image generation requests asynchronously in a batch.
Uses of Experimental in dev.langchain4j.model.mistralai

Classes in dev.langchain4j.model.mistralai with annotations of type Experimental

Modifier and Type

Class

Description

final class

MistralAiBatchChatModel

A BatchChatModel for the Mistral Batch API, which processes multiple chat requests asynchronously at 50% of the standard per-token price.
Uses of Experimental in dev.langchain4j.model.ollama

Classes in dev.langchain4j.model.ollama with annotations of type Experimental

Modifier and Type

Class

Description

class

OllamaDecisionModel

A DecisionModel backed by the decision models that Ollama runs locally, such as Nimble, Tev1 and Clef, through its System One API (POST /v1/systemone, Ollama v0.35.0 or later).

class

OllamaDecisionRequestParameters

The per-call parameters of an OllamaDecisionModel request.

class

OllamaImageModel

Experimental Ollama image generation model.
Uses of Experimental in dev.langchain4j.model.openai

Classes in dev.langchain4j.model.openai with annotations of type Experimental

Modifier and Type

Class

Description

class

OpenAiAudioTranscriptionModel

Represents an OpenAI audio model with a transcription interface, only gpt-4o-transcribe, gpt-4o-mini-transcribe, whisper-1 (which is powered by our open source Whisper V2 model), and gpt-4o-transcribe-diarize are supported.

enum

OpenAiAudioTranscriptionModelName

class

OpenAiDecisionModel

A DecisionModel backed by the OpenAI Decisions API (POST /v1/decisions), which answers yes/no, choice and scale questions about a text, image or conversation input with probabilities.

class

OpenAiDecisionRequestParameters

The per-call parameters of an OpenAiDecisionModel request.

class

OpenAiEmbeddingRequestParameters

OpenAI-specific EmbeddingRequestParameters, adding the parameters supported by the OpenAI embeddings API on top of the common modelName/dimensions.

class

OpenAiResponsesChatModel

class

OpenAiResponsesStreamingChatModel

class

OpenAiTextToSpeechModel

Represents an OpenAI text-to-speech model with a speech generation interface.

enum

OpenAiTextToSpeechModelName

Methods in dev.langchain4j.model.openai with annotations of type Experimental

Modifier and Type

Method

Description

CompletableFuture<ChatResponse>

OpenAiResponsesChatModel.doChatAsync(ChatRequest chatRequest)

OpenAiResponsesStreamingChatModel.Builder

OpenAiResponsesStreamingChatModel.Builder.streamingBufferSize(Integer streamingBufferSize)

Sets the size of the bounded back-pressure buffer for the reactive streaming path.

OpenAiStreamingChatModel.OpenAiStreamingChatModelBuilder

OpenAiStreamingChatModel.OpenAiStreamingChatModelBuilder.streamingBufferSize(Integer streamingBufferSize)

Sets the size of the bounded back-pressure buffer for the reactive (Flow.Publisher) streaming path.
Uses of Experimental in dev.langchain4j.model.openai.internal

Methods in dev.langchain4j.model.openai.internal with annotations of type Experimental

Modifier and Type

Method

Description

Flow.Publisher<ChatModelStreamingEvent>

OpenAiClient.chatCompletionPublisher(ChatCompletionRequest request, ChatCompletionOptions options)

default CompletableFuture<ParsedAndRawResponse<ResponseContent>>

SyncOrAsync.executeRawAsync()

B

OpenAiClient.Builder.streamingBufferSize(Integer streamingBufferSize)

Sets the size of the bounded back-pressure buffer used by the reactive streaming publisher (chatCompletionPublisher).
Uses of Experimental in dev.langchain4j.model.openaiofficial

Classes in dev.langchain4j.model.openaiofficial with annotations of type Experimental

Modifier and Type

Class

Description

final class

OpenAiOfficialBatchChatModel

Submits chat requests to the OpenAI Batch API, which processes them asynchronously at a reduced cost compared to real-time requests.

class

OpenAiOfficialChatRequestParameters

class

OpenAiOfficialChatResponseMetadata

class

OpenAiOfficialDecisionModel

A DecisionModel backed by the OpenAI Decisions API, using the official OpenAI Java SDK.

class

OpenAiOfficialDecisionRequestParameters

The per-call parameters of an OpenAiOfficialDecisionModel request.

class

OpenAiOfficialResponsesChatModel

ChatModel implementation using the official OpenAI Java client for the Responses API.

class

OpenAiOfficialResponsesChatRequestParameters

class

OpenAiOfficialResponsesChatResponseMetadata

class

OpenAiOfficialResponsesStreamingChatModel

StreamingChatModel implementation using the official OpenAI Java client for the Responses API.
Uses of Experimental in dev.langchain4j.model.scoring

Classes in dev.langchain4j.model.scoring with annotations of type Experimental

Modifier and Type

Class

Description

class

DecisionScoringModel

A ScoringModel backed by a DecisionModel: the score of a segment is the probability that the answer to "Does the document help answer the query?"

Methods in dev.langchain4j.model.scoring with annotations of type Experimental

Modifier and Type

Method

Description

default ScoringRequestParameters

ScoringModel.defaultRequestParameters()

The model's default per-call parameters, applied by ScoringModel.scoreAsync(ScoringRequest) to every request and overridden by any parameters set on the request itself.

default CompletableFuture<ScoringResponse>

ScoringModel.doScoreAsync(ScoringRequest request)

The provider hook behind ScoringModel.scoreAsync(ScoringRequest).

default CompletableFuture<ScoringResponse>

ScoringModel.scoreAsync(ScoringRequest request)

Scores the documents of the given request against its query without blocking the calling thread — the asynchronous/reactive counterpart of ScoringModel.scoreAll(List, String), used by the non-blocking RAG flow (see ReRankingContentAggregator).
Uses of Experimental in dev.langchain4j.model.scoring.request

Classes in dev.langchain4j.model.scoring.request with annotations of type Experimental

Modifier and Type

Class

Description

class

DefaultScoringRequestParameters

The default implementation of ScoringRequestParameters.

class

ScoringRequest

A request to score (re-rank) a list of ScoringRequest.documents() against a ScoringRequest.query().

interface

ScoringRequestParameters

The per-call parameters of a ScoringRequest, such as the ScoringRequestParameters.modelName() to use.
Uses of Experimental in dev.langchain4j.model.scoring.response

Classes in dev.langchain4j.model.scoring.response with annotations of type Experimental

Modifier and Type

Class

Description

class

ScoringResponse

The result of scoring a ScoringRequest: one relevance score per request document, in the same order, together with the ScoringResponseMetadata (model name, token/billing usage).

class

ScoringResponseMetadata

The metadata of a ScoringResponse: the ScoringResponseMetadata.modelName() that produced the scores and the ScoringResponseMetadata.tokenUsage() (or provider billing units) the call consumed.
Uses of Experimental in dev.langchain4j.model.typesafe

Classes in dev.langchain4j.model.typesafe with annotations of type Experimental

Modifier and Type

Class

Description

class

TypeSafeDecisionModel

A DecisionModel backed by the TypeSafe System One API (POST /v1/systemone), which serves decision models such as Jev.

class

TypeSafeDecisionRequestParameters

The per-call parameters of a TypeSafeDecisionModel request.
Uses of Experimental in dev.langchain4j.model.voyageai

Classes in dev.langchain4j.model.voyageai with annotations of type Experimental

Modifier and Type

Class

Description

class

VoyageAiEmbeddingRequestParameters

Voyage AI-specific EmbeddingRequestParameters, adding the parameters accepted by the Voyage AI embeddings API on top of the common inputType.
Uses of Experimental in dev.langchain4j.observability.api.event

Classes in dev.langchain4j.observability.api.event with annotations of type Experimental

Modifier and Type

Interface

Description

interface

AiServiceCancelledEvent

Invoked when an AI Service invocation is cancelled before it completes, for example when the Flow.Publisher subscription is cancelled, the returned CompletableFuture is cancelled, or StreamingHandle.cancel() is called from a TokenStream callback that receives a context, such as onPartialResponseWithContext.

enum

CompensationReason

Why a successfully-executed tool was compensated (rolled back).

interface

RetrievalAugmentedEvent

Invoked once per invocation of an AI Service that uses retrieval-augmented generation (RAG), configured with contentRetriever(...) or retrievalAugmentor(...), after the contents to add to the user message were retrieved, also when nothing was retrieved.

interface

ToolCompensatedEvent

Invoked after a successfully-executed tool is compensated (rolled back) - for example when another tool in the same round failed, or when the invocation was cancelled.

Methods in dev.langchain4j.observability.api.event with annotations of type Experimental

Modifier and Type

Method

Description

default boolean

ToolExecutedEvent.isError()

Returns whether the tool execution failed, for example because the tool threw an exception and the error was sent to the LLM.

Boolean

ToolExecutedEvent.ToolExecutedEventBuilder.isError()

Returns whether the tool execution failed, or null if it was not set.

ToolExecutedEvent.ToolExecutedEventBuilder

ToolExecutedEvent.ToolExecutedEventBuilder.isError(Boolean isError)

Sets whether the tool execution failed.

List<Content>

ToolCompensatedEvent.resultContents()

Returns the contents of the compensated tool's original (successful) result.

ToolCompensatedEvent.ToolCompensatedEventBuilder

ToolCompensatedEvent.ToolCompensatedEventBuilder.resultContents(List<Content> resultContents)

Sets the compensated tool's result contents.

List<Content>

ToolExecutedEvent.resultContents()

Returns the contents of the tool execution result.

ToolExecutedEvent.ToolExecutedEventBuilder

ToolExecutedEvent.ToolExecutedEventBuilder.resultContents(List<Content> resultContents)

Sets the tool execution result contents.
Uses of Experimental in dev.langchain4j.observability.api.listener

Classes in dev.langchain4j.observability.api.listener with annotations of type Experimental

Modifier and Type

Interface

Description

interface

AiServiceCancelledListener

A listener for AiServiceCancelledEvent, which occurs when an AI Service invocation is cancelled before it completes.

interface

RetrievalAugmentedListener

A listener for RetrievalAugmentedEvent, which occurs once per invocation of an AI Service that uses retrieval-augmented generation (RAG), configured with contentRetriever(...) or retrievalAugmentor(...), after the contents to add to the user message were retrieved.

interface

ToolCompensatedEventListener

A listener for ToolCompensatedEvent, which represents an event that occurs after a successfully-executed tool is compensated (rolled back) within an AI Service invocation.
Uses of Experimental in dev.langchain4j.observability.event

Classes in dev.langchain4j.observability.event with annotations of type Experimental

Modifier and Type

Class

Description

class

DefaultAiServiceCancelledEvent

Default implementation of AiServiceCancelledEvent.

class

DefaultRetrievalAugmentedEvent

Default implementation of RetrievalAugmentedEvent.

class

DefaultToolCompensatedEvent

Default implementation of ToolCompensatedEvent.

Methods in dev.langchain4j.observability.event with annotations of type Experimental

Modifier and Type

Method

Description

boolean

DefaultToolExecutedEvent.isError()

List<Content>

DefaultToolCompensatedEvent.resultContents()

List<Content>

DefaultToolExecutedEvent.resultContents()
Uses of Experimental in dev.langchain4j.observation.listener

Classes in dev.langchain4j.observation.listener with annotations of type Experimental

Modifier and Type

Class

Description

class

ObservationChatModelListener

Will use observations and micrometer metrics to generate telemetry based on the ChatModelListener lifecycle.
Uses of Experimental in dev.langchain4j.rag

Methods in dev.langchain4j.rag with annotations of type Experimental

Modifier and Type

Method

Description

default CompletableFuture<AugmentationResult>

RetrievalAugmentor.augmentAsync(AugmentationRequest augmentationRequest)

Non-blocking counterpart of RetrievalAugmentor.augment(AugmentationRequest), invoked by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service modes so the RAG flow never blocks the calling thread.

DefaultRetrievalAugmentor.DefaultRetrievalAugmentorBuilder

DefaultRetrievalAugmentor.DefaultRetrievalAugmentorBuilder.offloadBlocking(boolean offloadBlocking)

Controls what DefaultRetrievalAugmentor.augmentAsync(AugmentationRequest) does when a pipeline stage (query transformation, routing, retrieval or aggregation) is not genuinely asynchronous (its *Async method is not implemented).

Constructors in dev.langchain4j.rag with annotations of type Experimental

Modifier

Constructor

Description

DefaultRetrievalAugmentor(QueryTransformer queryTransformer, QueryRouter queryRouter, ContentAggregator contentAggregator, ContentInjector contentInjector, Executor executor, boolean offloadBlocking)
Uses of Experimental in dev.langchain4j.rag.content.aggregator

Methods in dev.langchain4j.rag.content.aggregator with annotations of type Experimental

Modifier and Type

Method

Description

default CompletableFuture<List<Content>>

ContentAggregator.aggregateAsync(Map<Query, Collection<List<Content>>> queryToContents)

Non-blocking counterpart of ContentAggregator.aggregate(Map), invoked by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service modes when RAG is configured.
Uses of Experimental in dev.langchain4j.rag.content.retriever

Methods in dev.langchain4j.rag.content.retriever with annotations of type Experimental

Modifier and Type

Method

Description

default ContentRetriever

ContentRetriever.addListener(ContentRetrieverListener listener)

Wraps this ContentRetriever with a listening retriever that dispatches events to the provided listener.

default ContentRetriever

ContentRetriever.addListeners(Collection<ContentRetrieverListener> listeners)

Wraps this ContentRetriever with a listening retriever that dispatches events to the provided listeners.

EmbeddingStoreContentRetriever.EmbeddingStoreContentRetrieverBuilder

EmbeddingStoreContentRetriever.EmbeddingStoreContentRetrieverBuilder.offloadBlocking(boolean offloadBlocking)

Controls what EmbeddingStoreContentRetriever.retrieveAsync(Query) does when the EmbeddingModel or EmbeddingStore is not genuinely asynchronous (does not implement doEmbedAsync/searchAsync).

default CompletableFuture<List<Content>>

ContentRetriever.retrieveAsync(Query query)

Non-blocking counterpart of ContentRetriever.retrieve(Query), invoked by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service modes when RAG is configured.
Uses of Experimental in dev.langchain4j.rag.content.retriever.hibernate

Classes in dev.langchain4j.rag.content.retriever.hibernate with annotations of type Experimental

Modifier and Type

Class

Description

class

HibernateContentRetriever

Using the SessionFactory and the ChatModel, this ContentRetriever attempts to generate and execute Hibernate queries for given natural language queries.

Constructors in dev.langchain4j.rag.content.retriever.hibernate with annotations of type Experimental

Modifier

Constructor

Description

HibernateContentRetriever(org.hibernate.SessionFactory sessionFactory, ChatModel chatModel, String databaseStructure, PromptTemplate promptTemplate, Integer maxRetries)

Creates an instance of a HibernateContentRetriever.
Uses of Experimental in dev.langchain4j.rag.content.retriever.listener

Classes in dev.langchain4j.rag.content.retriever.listener with annotations of type Experimental

Modifier and Type

Class

Description

class

ContentRetrieverErrorContext

The content retriever error context.

static class

ContentRetrieverErrorContext.Builder

Builder for ContentRetrieverErrorContext.

interface

ContentRetrieverListener

A ContentRetriever listener that listens for requests, responses and errors.

class

ContentRetrieverRequestContext

The content retriever request context.

static class

ContentRetrieverRequestContext.Builder

Builder for ContentRetrieverRequestContext.

class

ContentRetrieverResponseContext

The content retriever response context.

static class

ContentRetrieverResponseContext.Builder

Builder for ContentRetrieverResponseContext.
Uses of Experimental in dev.langchain4j.rag.query.router

Classes in dev.langchain4j.rag.query.router with annotations of type Experimental

Modifier and Type

Class

Description

class

DecisionModelQueryRouter

A QueryRouter that uses a DecisionModel to decide which ContentRetrievers can help answer the query.

Methods in dev.langchain4j.rag.query.router with annotations of type Experimental

Modifier and Type

Method

Description

default CompletableFuture<Collection<ContentRetriever>>

QueryRouter.routeAsync(Query query)

Non-blocking counterpart of QueryRouter.route(Query), invoked by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service modes when RAG is configured.
Uses of Experimental in dev.langchain4j.rag.query.transformer

Methods in dev.langchain4j.rag.query.transformer with annotations of type Experimental

Modifier and Type

Method

Description

default CompletableFuture<Collection<Query>>

QueryTransformer.transformAsync(Query query)

Non-blocking counterpart of QueryTransformer.transform(Query), invoked by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service modes when RAG is configured.
Uses of Experimental in dev.langchain4j.service

Classes in dev.langchain4j.service with annotations of type Experimental

Modifier and Type

Interface

Description

interface

AiServiceStreamingEvent

An event emitted by the reactive stream of a non-blocking streaming AI Service — that is, an AI Service method declared to return a Flow.Publisher of AiServiceStreamingEvent.

Methods in dev.langchain4j.service with annotations of type Experimental

Modifier and Type

Method

Description

default TokenStream

TokenStream.onPartialResponseWithContext(BiConsumer<PartialResponse, PartialResponseContext> handler)

The provided consumer will be invoked every time a new partial textual response (usually a single token) from a language model is available.

default TokenStream

TokenStream.onPartialThinking(Consumer<PartialThinking> partialThinkingHandler)

The provided consumer will be invoked every time a new partial thinking/reasoning text (usually a single token) from a language model is available.

default TokenStream

TokenStream.onPartialThinkingWithContext(BiConsumer<PartialThinking, PartialThinkingContext> handler)

The provided consumer will be invoked every time a new partial thinking/reasoning text (usually a single token) from a language model is available.

default TokenStream

TokenStream.onPartialToolCall(Consumer<PartialToolCall> partialToolCallHandler)

The provided consumer will be invoked every time a new partial tool call (usually containing a single token of the tool's arguments) from a language model is available.

default TokenStream

TokenStream.onPartialToolCallWithContext(BiConsumer<PartialToolCall, PartialToolCallContext> handler)

The provided consumer will be invoked every time a new partial tool call (usually containing a single token of the tool's arguments) from a language model is available.

default TokenStream

TokenStream.onUnmappedRawEvent(Consumer<Object> rawEventHandler)

The provided consumer will be invoked when a provider emits a raw streaming event that is not already exposed through one of the typed callbacks (such as TokenStream.onPartialResponse(Consumer), TokenStream.onPartialThinking(Consumer) or TokenStream.onToolExecuted(Consumer)).

AiServices<T>

AiServices.streamingBufferSize(int streamingBufferSize)

Sets the size of the bounded back-pressure buffer used by the reactive (Flow.Publisher) streaming path.
Uses of Experimental in dev.langchain4j.service.decision

Classes in dev.langchain4j.service.decision with annotations of type Experimental

Modifier and Type

Class

Description

final class

Choice<E>

The answer to a question of a decision service that chooses one of several options: the chosen option, the probability of each option and, if the model reports one, a confidence.

@interface

Decide

The question a decision model answers for a method of a decision service, or for a field of the object a method returns.

final class

DecisionResult<T>

The result of a decision service method together with the underlying DecisionResponse, which gives access to the raw answers (including probabilities), the model name and the token usage.

final class

DecisionServices

Creates implementations of Java interfaces whose methods are answered by a DecisionModel.

final class

Scale<E extends Enum<E>>

The answer to a question of a decision service that places the input on an ordered scale: the mean level, the probability of each level and, if the model reports one, a confidence.

final class

ThresholdContext

The question a ThresholdProvider is asked the threshold for.

interface

ThresholdProvider

Provides the threshold for boolean answers of a decision service: the answer is true when the probability of "yes" is greater than or equal to the threshold.
Uses of Experimental in dev.langchain4j.service.guardrail

Methods in dev.langchain4j.service.guardrail with annotations of type Experimental

Modifier and Type

Method

Description

default <MethodKey> CompletableFuture<UserMessage>

GuardrailService.executeGuardrailsAsync(MethodKey method, InputGuardrailRequest request)

Non-blocking counterpart of GuardrailService.executeGuardrails(Object, InputGuardrailRequest): runs the input guardrails without blocking the calling thread and yields the (possibly rewritten) UserMessage.

default <MethodKey, T> CompletableFuture<T>

GuardrailService.executeGuardrailsAsync(MethodKey method, OutputGuardrailRequest request)

Non-blocking counterpart of GuardrailService.executeGuardrails(Object, OutputGuardrailRequest) for the asynchronous (CompletableFuture) and reactive (Flow.Publisher) AI Service modes.

default <MethodKey> CompletableFuture<InputGuardrailResult>

GuardrailService.executeInputGuardrailsAsync(MethodKey method, InputGuardrailRequest request)

Non-blocking counterpart of GuardrailService.executeInputGuardrails(Object, InputGuardrailRequest).

default <MethodKey> CompletableFuture<OutputGuardrailResult>

GuardrailService.executeOutputGuardrailsAsync(MethodKey method, OutputGuardrailRequest request)

Non-blocking counterpart of GuardrailService.executeOutputGuardrails(Object, OutputGuardrailRequest).
Uses of Experimental in dev.langchain4j.service.tool

Classes in dev.langchain4j.service.tool with annotations of type Experimental

Modifier and Type

Class

Description

class

BeforeToolExecution

class

DecisionModelFilteringToolProvider

A ToolProvider that passes on only the tools of another tool provider that are relevant to the conversation, as decided by a DecisionModel.

Methods in dev.langchain4j.service.tool with annotations of type Experimental

Modifier and Type

Method

Description

default CompletableFuture<ToolExecutionResult>

ToolExecutor.executeAsync(ToolExecutionRequest request, InvocationContext context)

Non-blocking counterpart of ToolExecutor.executeWithContext(ToolExecutionRequest, InvocationContext), invoked by the asynchronous AI Service tool loop (AI Service methods returning CompletableFuture or CompletionStage), which composes the returned future instead of waiting on a thread.

List<Content>

ToolExecution.resultContents()

Returns the contents of the tool execution result.

ToolExecutionResult.Builder

ToolExecutionResult.Builder.resultContents(List<Content> resultContents)

Sets the contents of the tool execution result.

List<Content>

ToolExecutionResult.resultContents()

Returns the contents of the tool execution result that will be sent to the LLM.
Uses of Experimental in dev.langchain4j.service.tool.search

Classes in dev.langchain4j.service.tool.search with annotations of type Experimental

Modifier and Type

Class

Description

class

ToolSearchRequest

A request to search for tools.

class

ToolSearchResult

The result of a tool search.

interface

ToolSearchStrategy

Strategy for discovering tools dynamically by an LLM during an AI Service invocation.
Uses of Experimental in dev.langchain4j.service.tool.search.decision

Classes in dev.langchain4j.service.tool.search.decision with annotations of type Experimental

Modifier and Type

Class

Description

class

DecisionModelToolSearchStrategy

A ToolSearchStrategy that uses a DecisionModel to find the tools matching the search query of the LLM: for each searchable tool, the decision model answers whether the tool would help with the query (all tools in a single call), and the most relevant tools are returned.
Uses of Experimental in dev.langchain4j.service.tool.search.simple

Classes in dev.langchain4j.service.tool.search.simple with annotations of type Experimental

Modifier and Type

Class

Description

class

SimpleToolSearchStrategy

A ToolSearchStrategy that allows an LLM to search for available tools using simple case-insensitive contains matching.
Uses of Experimental in dev.langchain4j.service.tool.search.vector

Classes in dev.langchain4j.service.tool.search.vector with annotations of type Experimental

Modifier and Type

Class

Description

class

VectorToolSearchStrategy

A ToolSearchStrategy that uses vector similarity search to find relevant tools based on the semantic meaning of their names and descriptions.
Uses of Experimental in dev.langchain4j.skills

Classes in dev.langchain4j.skills with annotations of type Experimental

Modifier and Type

Class

Description

class

AbstractSkill

class

ActivateSkillToolConfig

class

ClassPathSkillLoader

Loads skills from the classpath (including cases when they are packaged inside a JAR file that is on the classpath).

class

DefaultFileSystemSkill

class

DefaultSkill

class

DefaultSkillResource

interface

FileSystemSkill

A Skill backed by the file system.

class

FileSystemSkillLoader

Loads skills from the file system.

class

ReadResourceToolConfig

interface

Skill

Represents a skill that can be used by an LLM.

interface

SkillResource

An additional resource associated with a Skill, such as a reference file, asset, or template that the LLM can read on demand.

class

Skills

Configures and exposes a set of Skills to an LLM.
Uses of Experimental in dev.langchain4j.skills.shell

Classes in dev.langchain4j.skills.shell with annotations of type Experimental

Modifier and Type

Class

Description

class

RunShellCommandToolConfig

class

ShellSkills

Configures and exposes a set of FileSystemSkills to an LLM using shell commands.
Uses of Experimental in dev.langchain4j.spi

Classes in dev.langchain4j.spi with annotations of type Experimental

Modifier and Type

Interface

Description

interface

CapturedContext

The context of a thread (tracing span, SLF4J MDC, security context, …), captured by ExecutorProvider.captureContext() so that it can be restored in tasks that run later, possibly on another thread.

interface

ExecutorProvider

SPI for supplying the Executor that LangChain4j uses to offload blocking work and run its asynchronous continuations — concurrent tool execution, blocking RAG / retrieval and embedding-store calls, moderation, in-process embedding models, retry backoff scheduling, and similar.
Uses of Experimental in dev.langchain4j.store.embedding

Methods in dev.langchain4j.store.embedding with annotations of type Experimental

Modifier and Type

Method

Description

default EmbeddingStore<Embedded>

EmbeddingStore.addListener(EmbeddingStoreListener listener)

Wraps this EmbeddingStore with a listening store that dispatches events to the provided listener.

default EmbeddingStore<Embedded>

EmbeddingStore.addListeners(List<EmbeddingStoreListener> listeners)

Wraps this EmbeddingStore with a listening store that dispatches events to the provided listeners.

default CompletableFuture<EmbeddingSearchResult<Embedded>>

EmbeddingStore.searchAsync(EmbeddingSearchRequest request)

Non-blocking counterpart of EmbeddingStore.search(EmbeddingSearchRequest), used by the asynchronous and reactive RAG flow (see EmbeddingStoreContentRetriever.retrieveAsync).
Uses of Experimental in dev.langchain4j.store.embedding.filter.builder.sql

Classes in dev.langchain4j.store.embedding.filter.builder.sql with annotations of type Experimental

Modifier and Type

Class

Description

class

ColumnDefinition

class

LanguageModelSqlFilterBuilder

Given a natural language Query, this class creates a suitable Filter using a language model.

class

TableDefinition
Uses of Experimental in dev.langchain4j.store.embedding.filter.parser.sql

Classes in dev.langchain4j.store.embedding.filter.parser.sql with annotations of type Experimental

Modifier and Type

Class

Description

class

SqlFilterParser

Parses an SQL "WHERE" clause into a Filter object using JSqlParser.
Uses of Experimental in dev.langchain4j.store.embedding.listener

Classes in dev.langchain4j.store.embedding.listener with annotations of type Experimental

Modifier and Type

Class

Description

class

EmbeddingStoreErrorContext<Embedded>

The embedding store error context.

interface

EmbeddingStoreListener

A EmbeddingStore listener that listens for requests, responses and errors.

class

EmbeddingStoreRequestContext<Embedded>

The embedding store request context.

static final class

EmbeddingStoreRequestContext.Add<Embedded>

The add(...) request context.

static final class

EmbeddingStoreRequestContext.AddAll<Embedded>

The addAll(...) request context.

static final class

EmbeddingStoreRequestContext.Remove<Embedded>

The remove(String) request context.

static final class

EmbeddingStoreRequestContext.RemoveAll<Embedded>

The removeAll() request context.

static final class

EmbeddingStoreRequestContext.RemoveAllFilter<Embedded>

The removeAll(Filter) request context.

static final class

EmbeddingStoreRequestContext.RemoveAllIds<Embedded>

The removeAll(ids) request context.

static final class

EmbeddingStoreRequestContext.Search<Embedded>

The search(...) request context.

class

EmbeddingStoreResponseContext<Embedded>

The embedding store response context.

static final class

EmbeddingStoreResponseContext.Add<Embedded>

The add(...) response context.

static final class

EmbeddingStoreResponseContext.AddAll<Embedded>

The addAll(...) response context.

static final class

EmbeddingStoreResponseContext.Remove<Embedded>

The remove(String) response context.

static final class

EmbeddingStoreResponseContext.RemoveAll<Embedded>

The removeAll() response context.

static final class

EmbeddingStoreResponseContext.RemoveAllFilter<Embedded>

The removeAll(Filter) response context.

static final class

EmbeddingStoreResponseContext.RemoveAllIds<Embedded>

The removeAll(ids) response context.

static final class

EmbeddingStoreResponseContext.Search<Embedded>

The search(...) response context.
Uses of Experimental in dev.langchain4j.store.memory.chat

Methods in dev.langchain4j.store.memory.chat with annotations of type Experimental

Modifier and Type

Method

Description

default CompletableFuture<Void>

ChatMemoryStore.deleteMessagesAsync(Object memoryId)

Non-blocking counterpart of ChatMemoryStore.deleteMessages(Object), used by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service APIs.

default CompletableFuture<List<ChatMessage>>

ChatMemoryStore.getMessagesAsync(Object memoryId)

Non-blocking counterpart of ChatMemoryStore.getMessages(Object), used by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service APIs.

default CompletableFuture<Void>

ChatMemoryStore.updateMessagesAsync(Object memoryId, List<ChatMessage> messages)

Non-blocking counterpart of ChatMemoryStore.updateMessages(Object, List), used by the asynchronous (CompletableFuture/CompletionStage) and reactive (Flow.Publisher) AI Service APIs.
Uses of Experimental in dev.langchain4j.web.search

Methods in dev.langchain4j.web.search with annotations of type Experimental

Modifier and Type

Method

Description

default CompletableFuture<WebSearchResults>

WebSearchEngine.searchAsync(WebSearchRequest webSearchRequest)

Non-blocking counterpart of WebSearchEngine.search(WebSearchRequest), used by the asynchronous and reactive RAG flow (see WebSearchContentRetriever).

Uses of Annotation Interfacedev.langchain4j.Experimental

Uses of Experimental in dev.langchain4j.agent.tool

Uses of Experimental in dev.langchain4j.agentic.patterns.cascade

Uses of Experimental in dev.langchain4j.agentic.patterns.decisionrouter

Uses of Experimental in dev.langchain4j.data.message

Uses of Experimental in dev.langchain4j.exception

Uses of Experimental in dev.langchain4j.experimental.rag.content.retriever.sql

Uses of Experimental in dev.langchain4j.guardrail

Uses of Experimental in dev.langchain4j.guardrails

Uses of Experimental in dev.langchain4j.http.client

Uses of Experimental in dev.langchain4j.http.client.apache

Uses of Experimental in dev.langchain4j.http.client.jdk

Uses of Experimental in dev.langchain4j.http.client.okhttp

Uses of Experimental in dev.langchain4j.http.client.sse

Uses of Experimental in dev.langchain4j.invocation

Uses of Experimental in dev.langchain4j.mcp.client

Uses of Experimental in dev.langchain4j.memory

Uses of Experimental in dev.langchain4j.micrometer.metrics.listeners

Uses of Experimental in dev.langchain4j.model.anthropic

Uses of Experimental in dev.langchain4j.model.audio

Uses of Experimental in dev.langchain4j.model.azure

Uses of Experimental in dev.langchain4j.model.batch

Uses of Experimental in dev.langchain4j.model.bedrock

Uses of Experimental in dev.langchain4j.model.catalog

Uses of Experimental in dev.langchain4j.model.chat

Uses of Experimental in dev.langchain4j.model.chat.response

Uses of Experimental in dev.langchain4j.model.chat.router

Uses of Experimental in dev.langchain4j.model.decision

Uses of Experimental in dev.langchain4j.model.decision.listener

Uses of Experimental in dev.langchain4j.model.decision.request

Uses of Experimental in dev.langchain4j.model.decision.response

Uses of Experimental in dev.langchain4j.model.embedding

Uses of Experimental in dev.langchain4j.model.embedding.listener

Uses of Experimental in dev.langchain4j.model.embedding.request

Uses of Experimental in dev.langchain4j.model.embedding.response

Uses of Experimental in dev.langchain4j.model.google.genai

Uses of Experimental in dev.langchain4j.model.googleai

Uses of Experimental in dev.langchain4j.model.image

Uses of Experimental in dev.langchain4j.model.mistralai

Uses of Experimental in dev.langchain4j.model.ollama

Uses of Experimental in dev.langchain4j.model.openai

Uses of Experimental in dev.langchain4j.model.openai.internal

Uses of Experimental in dev.langchain4j.model.openaiofficial

Uses of Experimental in dev.langchain4j.model.scoring

Uses of Experimental in dev.langchain4j.model.scoring.request

Uses of Experimental in dev.langchain4j.model.scoring.response

Uses of Experimental in dev.langchain4j.model.typesafe

Uses of Experimental in dev.langchain4j.model.voyageai

Uses of Experimental in dev.langchain4j.observability.api.event

Uses of Experimental in dev.langchain4j.observability.api.listener

Uses of Experimental in dev.langchain4j.observability.event

Uses of Experimental in dev.langchain4j.observation.listener

Uses of Experimental in dev.langchain4j.rag

Uses of Experimental in dev.langchain4j.rag.content.aggregator

Uses of Experimental in dev.langchain4j.rag.content.retriever

Uses of Experimental in dev.langchain4j.rag.content.retriever.hibernate

Uses of Experimental in dev.langchain4j.rag.content.retriever.listener

Uses of Experimental in dev.langchain4j.rag.query.router

Uses of Experimental in dev.langchain4j.rag.query.transformer

Uses of Experimental in dev.langchain4j.service

Uses of Experimental in dev.langchain4j.service.decision

Uses of Experimental in dev.langchain4j.service.guardrail

Uses of Experimental in dev.langchain4j.service.tool

Uses of Experimental in dev.langchain4j.service.tool.search

Uses of Experimental in dev.langchain4j.service.tool.search.decision

Uses of Experimental in dev.langchain4j.service.tool.search.simple

Uses of Experimental in dev.langchain4j.service.tool.search.vector

Uses of Experimental in dev.langchain4j.skills

Uses of Experimental in dev.langchain4j.skills.shell

Uses of Experimental in dev.langchain4j.spi

Uses of Experimental in dev.langchain4j.store.embedding

Uses of Experimental in dev.langchain4j.store.embedding.filter.builder.sql

Uses of Experimental in dev.langchain4j.store.embedding.filter.parser.sql

Uses of Experimental in dev.langchain4j.store.embedding.listener

Uses of Experimental in dev.langchain4j.store.memory.chat

Uses of Experimental in dev.langchain4j.web.search

Uses of Annotation Interface
dev.langchain4j.Experimental