Table of Contents

Namespace GenerativeAI.Types

Request and response types (DTOs) for the Gemini API and Vertex AI: content, parts, generation and tool configuration, the Live API messages (LiveClientMessage, LiveServerMessage), batch, caching, files, File Search, Interactions, tuning, webhooks and agent API types. Names and JSON property names follow the official REST API and python-genai.

Classes

ActivityEnd

Marks the end of user activity. This can only be sent if automatic (i.e. server-side) activity detection is disabled.

ActivityStart

Marks the start of user activity. This can only be sent if automatic (i.e. server-side) activity detection is disabled.

Agent

A managed agent (/agents): a reusable agent definition (base agent, configuration, system instruction, tools and base environment) that interactions can reference by ID.

AgentApisSerializerContext

Source-generated JSON metadata for the agent-related API types (Voices, Agents, Environments, Triggers, Credentials and Decisions). The clients of these APIs put it first in their resolver chain (followed by InteractionsSerializerContext for the shared Interactions types), so they work with Native AOT and trimming.

AgentEnvironment

A sandbox environment (/environments) used by agent interactions, with its mounted sources and workspace files. Named AgentEnvironment to avoid clashing with Environment.

AntigravityAgentConfig

Configuration of the Antigravity agent (antigravity).

ApiKeyConfig

Config for authentication with API key.

AsyncBatchEmbedContentRequest

Request for batch embedding generation.

AttributionSourceId

Identifier for the source contributing to this attribution.

AudioChunk

Representation of an audio chunk.

AudioResponseFormat

Format of audio output.

AudioTranscriptionConfig

Configuration settings for audio transcription in multimodal live sessions.

AuthConfig

Auth configuration to run the extension.

AuthConfigGoogleServiceAccountConfig

Config for Google Service Account Authentication.

AuthConfigHttpBasicAuthConfig

Config for HTTP Basic Authentication.

AuthConfigOauthConfig

Config for user oauth.

AuthConfigOidcConfig

Config for user OIDC auth.

AuthToken

An ephemeral auth token for the Live API. Pass Name as the API key of a GoogleAIPlatformAdapter used by a Live client; the client then connects to the constrained endpoint.

AutoRoutingMode

When automated routing is specified, the routing will be determined by the pretrained routing model and customer provided model routing preference.

AutomaticActivityDetection

Configures automatic detection of activity.

AutoraterConfig

Autorater config used for model evaluation during tuning.

AvatarConfig

Configures the avatar to be used in the live session.

BatchBigQueryDestination

BigQuery destination for batch output.

BatchBigQuerySource

BigQuery source for batch input.

BatchConfig

Batch configuration containing input config.

BatchContentOutput

Output wrapper for content generation batch jobs.

BatchContentResponsesWrapper

Wrapper for content generation batch responses (nested structure).

BatchCreateChunksRequest

Represents a request to batch create chunks.

BatchCreateChunksResponse

Represents the response from a batch create chunks operation.

BatchDeleteChunksRequest

Represents a request to batch delete chunks.

BatchEmbedContentRequest

Represents a batched request for embedding content.

BatchEmbedContentsResponse

Represents the response to a BatchEmbedContentRequest.

BatchEmbeddingOutput

Output wrapper for embedding batch jobs.

BatchEmbeddingResponsesWrapper

Wrapper for embedding batch responses (nested structure).

BatchGcsDestination

GCS destination for batch output.

BatchGcsSource

GCS source for batch input.

BatchGenerateContentRequest

Request for batch content generation.

BatchInputConfig

Input configuration for a batch job (Google AI / MLDev format).

BatchJob

Represents a batch processing job for generating content or embeddings at scale. Batch jobs allow processing large numbers of requests asynchronously.

BatchJobDestination

Configuration for the output data destination of a batch job.

BatchJobResponse

Response wrapper for batch job operations (Google AI format).

BatchJobResponseConverter

Custom JSON converter for BatchJobResponse that handles dynamic response parsing based on the metadata type field.

BatchJobSource

Configuration for the input data source of a batch job.

BatchOutputInfo

Information about the output of a batch prediction job (Vertex AI only).

BatchStats

Statistics about a batch job's progress.

BatchUpdateChunksRequest

Represents the request body for batch updating chunks.

BatchUpdateChunksResponse

Represents the response body for batch updating chunks.

BatchWrapper

Wrapper for batch request (Google AI / MLDev format).

BearerTokenCredentialConfig

An HTTP bearer token (bearer_token).

Blob

Raw media bytes. Text should not be sent as raw bytes, use the 'text' field.

CachedContent

Content that has been preprocessed and can be used in subsequent requests to GenerativeService. Cached content can be only used with the model it was created for.

Candidate

A response candidate generated from the model.

ChatSessionBackUpData

Represents a backup object for a chat session containing session-related data and configurations.

Checkpoint

Describes the machine learning model version checkpoint.

ChoiceDecisionAnswer

The answer to a ChoiceDecisionQuestion (choice).

ChoiceDecisionQuestion

A multiple-choice question (choice).

Chunk

A Chunk is a subpart of a Document that is treated as an independent unit for the purposes of vector representation and storage. A Corpus can have a maximum of 1 million Chunks.

ChunkData

Extracted data that represents the Chunk content.

ChunkingConfig

Tells the service how to split a file into chunks when adding it to a FileSearchStore.

CitationMetadata

A collection of source attributions for a piece of content.

CitationSource

A citation to a source for a portion of a specific response.

CodeExecutionResult

Result of executing the ExecutableCode. Only generated when using the CodeExecutionTool, and always follows a Part containing the ExecutableCode.

CodeExecutionTool

Tool that executes code generated by the model, and automatically returns the result to the model. See also ExecutableCode and CodeExecutionResult which are only generated when using this tool.

CodeMenderAgentConfig

Configuration of the CodeMender agent (code-mender).

CodeMenderFindRequest

Parameters of a CodeMender FIND session.

CodeMenderFixRequest

Parameters of a CodeMender FIX session.

CodeMenderSessionConfig

CodeMender session configuration.

CodeMenderSourceFile

A source file given to the CodeMender agent.

CompositeReinforcementTuningRewardConfig

A composite (weighted) reward function configuration for reinforcement tuning.

CompositeReinforcementTuningRewardConfigWeightedRewardConfig

A weighted single reward configuration inside a CompositeReinforcementTuningRewardConfig.

ComputeTokensRequest

Request body of the Vertex AI computeTokens method.

ComputeTokensResponse

Response of the Vertex AI computeTokens method.

ComputerUse

Tool to support computer use.

Condition

Filter condition applicable to a single key.

Content

The base structured datatype containing multi-part content of a message. A Content includes a Role field designating the producer of the Content and a Parts field containing multi-part data that contains the content of the message turn.

ContentEmbedding

Represents a list of floats that make up an embedding.

ContentEmbeddingStatistics

Statistics of the input text associated with the result of content embedding.

ContextUpdateTools

A wrapper around the list of tools in a LiveClientContextUpdate. The wrapper distinguishes "no tools update" (wrapper unset) from "clear all tools" (wrapper set with an empty list).

ContextWindowCompressionConfig

Configures context window compression mechanism. If included, server will compress context window to fit into given length.

Corpus

A Corpus is a collection of Document. A project can create up to 5 corpora.

CountTokensRequest

Request to count tokens.

CountTokensResponse

Response to count tokens.

CreateAuthTokenConfig

Options for creating an ephemeral Live API auth token.

CreateAuthTokenRequest

Request body for creating an ephemeral auth token (POST v1alpha/auth_tokens). Most callers should use CreateAuthTokenConfig, which builds the field mask for you.

CreateChunkRequest

Represents a request to create a Chunk.

CreateDecisionRequest

Request body of POST /decisions: evaluates one or more structured decision questions (NOUL, choice or score) against a shared Input.

CreateDecisionResponse

Response of POST /decisions.

CreateEnvironmentRequest

Request body of POST /environments.

CreateInteractionRequest

Request body of POST /interactions. Set either Model (model interaction) or Agent (agent interaction, e.g. Deep Research or Antigravity), not both.

CreateTriggerRequest

Request body of POST /triggers.

CreateTuningJobConfig

Optional configuration for TuneAsync(string, TuningDataset, CreateTuningJobConfig?, CancellationToken), mirroring CreateTuningJobConfig of the google-genai SDK.

CreateVoiceRequest

Request body of POST /voices.

CreateWebhookRequest

The request body of POST /webhooks (python-genai WebhookInput).

Credential

A stored credential (/credentials) that agent environments can use without exposing the secret (e.g. as an environment variable or a bearer token injected into allow-listed egress requests). Secrets are write-only and never returned.

CredentialConfig

The body of POST /credentials (create, Id required) and PATCH /credentials/{id} (update, Id not sent), discriminated by type (see CredentialTypes). For updates, only the fields that are set are sent; combine with an update_mask to clear fields.

CredentialInjectionLocations

Known values of InjectionLocation.

CredentialStatuses

Known credential statuses.

CredentialTypes

Known credential types.

CustomMetadata

User provided metadata stored as key-value pairs.

CustomizedAvatar

Configures the customized avatar to be used in the live session.

Dataset

Dataset for training or validation.

DecisionAnswer

The answer to a decision question, discriminated by type. Unknown types are read as UnknownDecisionAnswer.

DecisionBlockReasons

Known values of BlockReason.

DecisionModels

Models of the Decisions API.

DecisionPromptFeedback

Feedback on the input of a decision request.

DecisionQuestion

A decision question, discriminated by type (see DecisionQuestionTypes). Unknown types are read as UnknownDecisionQuestion.

DecisionQuestionTypes

Known decision question and answer types.

DeepResearchAgentConfig

Configuration of the Deep Research agent (deep-research).

DeleteChunkRequest

Represents a request to delete a specific chunk.

DeleteVoiceResponse

Response of DELETE /voices/{id}.

DistillationHyperParameters

Hyperparameters for distillation.

DistillationSamplingSpec

Spec for the teacher model sampling of a distillation job.

DistillationSpec

Tuning spec for distillation.

Document

A Document is a collection of Chunks. A Corpus can have a maximum of 10,000 Documents.

Duration

A Duration represents a signed, fixed-length span of time represented as a count of seconds and fractions of seconds at nanosecond resolution.

DurationJsonConverter

JSON converter for Duration.

DynamicAgentConfig

Free-form agent configuration (dynamic); put the settings in ExtensionData.

DynamicRetrievalConfig

Describes the options to customize dynamic retrieval.

EmbedContentRequest

Represents a request containing the content for the model to embed.

EmbedContentResponse

Represents the response to an EmbedContentRequest.

EmbedRequestBody

The actual embedding request body for batch processing.

EmbedRequestWrapper

Wrapper for a single embed request in batch processing.

EmbedRequestsWrapper

Wrapper for embedding requests array (Google AI nested structure).

EmbeddingBatchConfig

Batch configuration for embeddings.

EmbeddingBatchInputConfig

Input configuration for embedding batch jobs (Google AI format).

EmbeddingBatchWrapper

Wrapper for embedding batch request (Google AI format).

EncryptionSpec

Represents a customer-managed encryption key spec that can be applied to a top-level resource.

Endpoint

An endpoint where you deploy models.

EnterpriseWebSearch

Tool to search public web data, powered by Vertex AI Search and Sec4 compliance.

EnvironmentFile

A file or directory in an environment workspace.

EnvironmentFileTypes

Known environment file entry types.

EnvironmentNetworkModes

Known environment network modes.

EnvironmentSourceTypes

Known environment source types (Type).

EnvironmentStatuses

Known environment statuses.

EnvironmentVariableCredentialConfig

A secret exposed to the environment as a variable (environment_variable).

EvaluateDatasetRun

An evaluation run of a tuning job.

EvaluationConfig

Evaluation configuration for model tuning. Defines how the tuned model should be evaluated during and after training.

ExecutableCode

Code generated by the model that is meant to be executed, and the result returned to the model. Only generated when using the CodeExecutionTool tool, in which the code will be automatically executed, and a corresponding CodeExecutionResult will also be generated.

ExternalApi

Retrieve from data source powered by external API for grounding. The external API is not owned by Google, but needs to follow the pre-defined API spec.

ExternalApiElasticSearchParams

The search parameters to use for the ELASTIC_SEARCH spec.

ExternalApiSimpleSearchParams

The search parameters to use for SIMPLE_SEARCH spec.

FileData

URI based data.

FileSearch

Tool that retrieves relevant content from File Search stores (managed retrieval-augmented generation).

FileSearchDocument

A document inside a FileSearchStore: a file that has been chunked and embedded for retrieval.

FileSearchOperation

Long-running operation returned when a file is uploaded to (uploadToFileSearchStore) or imported into (importFile) a FileSearchStore.

FileSearchOperationResponse

The result of a completed upload or import into a FileSearchStore.

FileSearchSerializerContext

Source-generated serialization metadata for the File Search stores and documents types.

FileSearchStore

A File Search store: a managed collection of FileSearchDocuments that the FileSearch tool can retrieve from (managed retrieval-augmented generation).

FullFineTuningSpec

Tuning spec for full fine tuning.

FunctionCall

A predicted FunctionCall returned from the model that contains a string representing the Name with the arguments and their values.

FunctionCallingConfig

Configuration for specifying function calling behavior.

FunctionDeclaration

Declaration of a function that can be used as a Tool.

FunctionResponse

The result output from a FunctionCall that contains a string representing the Name and a structured JSON object containing any output from the function is used as context to the model. This should contain the result of a FunctionCall made based on model prediction.

FunctionResponseBlob

Raw media bytes for function response. Text should not be sent as raw bytes, use the FunctionResponse.response field.

FunctionResponseFileData

URI based data for function response.

FunctionResponsePart

A datatype containing media that is part of a FunctionResponse message. A FunctionResponsePart consists of data which has an associated datatype. A FunctionResponsePart can only contain one of the accepted types in FunctionResponsePart.data. A FunctionResponsePart must have a fixed IANA MIME type identifying the type and subtype of the media if the inline_data field is filled with raw bytes.

GcsDestination

The Google Cloud Storage location where the output is to be written to.

GenerateAnswerRequest

Request to generate an answer.

GenerateAnswerResponse

Response from the model for a grounded answer.

GenerateContentRequest

Request to generate content.

GenerateContentRequestForCountToken

Request to generate a completion from the model.

GenerateContentResponse

Response from the model supporting multiple candidate responses. Safety ratings and content filtering are reported for both prompt in PromptFeedback and for each candidate in FinishReason and in SafetyRatings. The API:

  • Returns either all requested candidates or none of them
  • Returns no candidates at all only if there was something wrong with the prompt (check PromptFeedback)
  • Reports feedback on each candidate in FinishReason and SafetyRatings.
GenerateImageRequest

Represents the request for generating images.

GenerateImageResponse

Represents the response from the image generation API.

GenerateVideosConfig

Configuration settings for a video generation request.

GenerateVideosOperation

Represents a long-running operation, specifically for video generation tasks. Poll this resource to check the status and retrieve results upon completion.

GenerateVideosRequest

Represents the parameters sent when requesting video generation.

GenerateVideosResponse

The response returned after a successful video generation request (often part of an Operation).

GeneratedVideo

Represents a single generated video as part of a video generation response.

GenerationConfig

Configuration options for model generation and outputs. Not all parameters are configurable for every model.

GenerationOutputConfig

Where to write generated output. Vertex AI only.

GoogleAIListBatchJobsResponse

Response for the list batch jobs method (Google AI wire format). Contains operations array that needs to be transformed to batch jobs.

GoogleAITunedModel

A tuned model resource of the Gemini Developer API (tunedModels/{id}).

GoogleLongRunningListOperationsResponse

The response message for Operations.ListOperations.

GoogleLongRunningOperation

This resource represents a long-running operation that is the result of a network API call.

GoogleMaps

Tool to support Google Maps in Model.

GoogleMapsGroundingTypes

The kinds of Google Maps grounding to use. Vertex AI only.

GoogleMapsPlaces

Places grounding settings. Currently has no options; set an instance to enable it.

GoogleMapsRouting

Routing grounding settings. Currently has no options; set an instance to enable it.

GoogleSchemaHelper

Provides helper methods for converting JSON schemas to Google-compatible formats.

GoogleSearchRetrievalTool

Tool to retrieve public web data for grounding, powered by Google.

GoogleSearchTool

GoogleSearch tool type. Tool to support Google Search in Model. Powered by Google.

GoogleTypeDate

Represents a whole or partial calendar date, such as a birthday.

GroundingAttribution

Attribution for a source that contributed to an answer.

GroundingChunk

Grounding chunk.

GroundingChunkCustomMetadata

A custom metadata entry attached to a File Search document.

GroundingChunkImage

An image used to ground the response (from Google Search image search).

GroundingChunkMaps

Chunk from Google Maps.

GroundingChunkMapsRoute

A route returned by Google Maps grounding.

GroundingChunkRetrievedContext

Chunk from context retrieved by the retrieval tools.

GroundingChunkStringList

A list of string values in custom metadata.

GroundingMetadata

Metadata returned to client when grounding is enabled.

GroundingMetadataSourceFlaggingUri

Source content flagging uri for a place or review. This is currently populated only for Google Maps grounding.

GroundingPassage

Passage included inline with a grounding configuration.

GroundingPassageId

Identifier for a part within a GroundingPassage.

GroundingPassages

A repeated list of passages.

GroundingSource

The sources in which to ground the answer.

GroundingSupport

Grounding support.

HistoryConfig

Configuration for history exchange between client and server.

Hyperparameters

Hyperparameters controlling the tuning process. Read more at https://ai.google.dev/docs/model_tuning_guidance

ImageCaptioningParameters

Represents the parameters for the Image Captioning API.

ImageCaptioningRequest

Represents the request body for the Image Captioning API.

ImageCaptioningResponse

Represents the response body from the Image Captioning API.

ImageConfig

Configuration for image generation in models that support image output (e.g., gemini-2.5-flash-image, gemini-3-pro-image-preview).

ImageData

Represents the image data, which can be either base64 encoded or a Cloud Storage URI.

ImageGenerationInstance

Represents an instance for image generation.

ImageGenerationParameters

Represents the parameters for image generation.

ImageInstance

Represents an image instance in the ImageCaptioningRequest.

ImageOutputOptions

The image output format for generated images.

ImageResponseFormat

Format of image output.

ImageSample

Represents an image, which can be provided via Google Cloud Storage URI or raw byte data.

ImageSearch

Image search settings. Currently has no options; set an instance to enable image search.

ImageSource

Represents the image source, either as base64 encoded bytes or a Cloud Storage URI.

ImportFileConfig

Optional settings for importing an existing Files API file into a FileSearchStore.

ImportFileRequest

Request body for fileSearchStores.importFile.

InlinedEmbedContentResponse

Represents an inlined embedding response from batch processing.

InlinedRequest

Represents an inlined request for batch processing without external storage.

InlinedRequestBody

The actual request body for batch inlined requests.

InlinedRequestWrapper

Wrapper for inlined request in batch processing (Google AI format).

InlinedResponse

Represents an inlined response from batch processing.

InputFeedback

Feedback related to the input data used to answer the question, as opposed to the model-generated response to the question.

Interaction

The Interaction resource returned by the Interactions API (/interactions). The output is exposed as Steps; OutputText, OutputImage, OutputAudio and OutputVideo are convenience accessors computed from the steps, like python-genai's helpers.

InteractionAgentConfig

Agent configuration (agent_config), discriminated by type.

InteractionAgents

Agent identifiers listed by the Interactions API AgentOption enum (python-genai). The API also accepts other agent IDs.

InteractionAllowedTools

Tools allowed for an MCP server or a tool choice.

InteractionAnnotation

A citation or other annotation attached to text content, discriminated by type.

InteractionArgumentsDelta

A chunk of function-call arguments JSON (arguments_delta).

InteractionAudioContent

An audio content block.

InteractionAudioDelta

An audio delta (audio).

InteractionAudioResponseFormat

Audio output format (type: audio).

InteractionCallStep

Base class for tool call steps, which carry the call Id.

InteractionCodeExecutionArguments

Arguments of a code execution call.

InteractionCodeExecutionCallDelta

A code execution call delta (code_execution_call).

InteractionCodeExecutionCallStep

A code execution call (code_execution_call).

InteractionCodeExecutionResultDelta

A code execution result delta (code_execution_result).

InteractionCodeExecutionResultStep

The result of a code execution (code_execution_result).

InteractionCodeExecutionTool

The code execution tool (code_execution).

InteractionCompletedEvent

The interaction completed (interaction.completed).

InteractionComputerUseTool

The computer use tool (computer_use).

InteractionContent

A content block of the Interactions API (text, image, audio, video or document), discriminated by type. Unknown content types are read as UnknownInteractionContent.

InteractionContentTypes

Known values of the type discriminator of InteractionContent.

InteractionCreatedEvent

The interaction was created (interaction.created).

InteractionDeliveryModes

Known values for the delivery field of media response formats.

InteractionDeltaTypes

Known values of the type discriminator of InteractionStepDeltaData.

InteractionDocumentContent

A document content block (PDF, CSV, Office documents, ...).

InteractionDocumentDelta

A document delta (document).

InteractionEnvironment

A remote sandbox environment configuration for agent interactions.

InteractionEnvironmentReference

The environment of an interaction: either an existing environment ID or an InteractionEnvironment configuration.

InteractionEnvironmentSource

A source mounted into a remote agent environment.

InteractionError

An error recorded on an interaction or sent in an error event.

InteractionErrorEvent

An error occurred while streaming (error).

InteractionEventTypes

Known values of the event_type discriminator of InteractionSseEvent.

InteractionExternalSearchConfig

Configuration of a third-party search provider (Exa.ai or Parallel.ai) for the retrieval tool.

InteractionFileCitation

A file citation (file_citation).

InteractionFileSearchCallDelta

A file search call delta (file_search_call).

InteractionFileSearchCallStep

A file search call (file_search_call).

InteractionFileSearchResultDelta

A file search result delta (file_search_result).

InteractionFileSearchResultStep

The result of a file search call (file_search_result).

InteractionFileSearchTool

The file search tool (file_search).

InteractionFunctionCallStep

A function call requested by the model (function_call).

InteractionFunctionResultDelta

A function result delta (function_result).

InteractionFunctionResultStep

The result of a function call, sent back by the client (function_result).

InteractionFunctionTool

A function declaration (function).

InteractionGenerationConfig

Generation parameters for model interactions (generation_config). Not allowed together with an agent.

InteractionGoogleMapsCallDelta

A Google Maps call delta (google_maps_call).

InteractionGoogleMapsCallStep

A Google Maps call (google_maps_call).

InteractionGoogleMapsPlace

A place returned by Google Maps.

InteractionGoogleMapsResult

One Google Maps result.

InteractionGoogleMapsResultDelta

A Google Maps result delta (google_maps_result).

InteractionGoogleMapsResultStep

The result of a Google Maps call (google_maps_result).

InteractionGoogleMapsTool

The Google Maps grounding tool (google_maps).

InteractionGoogleSearchCallDelta

A Google Search call delta (google_search_call).

InteractionGoogleSearchCallStep

A Google Search call (google_search_call).

InteractionGoogleSearchResult

One Google Search result.

InteractionGoogleSearchResultDelta

A Google Search result delta (google_search_result).

InteractionGoogleSearchResultStep

The result of a Google Search call (google_search_result).

InteractionGoogleSearchTool

The Google Search grounding tool (google_search).

InteractionGroundingToolCount

Usage count of a grounding tool.

InteractionImageConfig

Image generation configuration (deprecated in the API).

InteractionImageContent

An image content block.

InteractionImageDelta

An image delta (image).

InteractionImageResponseFormat

Image output format (type: image).

InteractionInput

The input of an interaction: a string, a single content block, a list of content blocks, or a list of steps (for example the previous turns plus function results).

InteractionMcpServerTool

A remote MCP server whose tools the model may call (mcp_server).

InteractionMcpServerToolCallDelta

An MCP server tool call delta (mcp_server_tool_call).

InteractionMcpServerToolCallStep

A call to a tool of a remote MCP server (mcp_server_tool_call).

InteractionMcpServerToolResultDelta

An MCP server tool result delta (mcp_server_tool_result).

InteractionMcpServerToolResultStep

The result of an MCP server tool call (mcp_server_tool_result).

InteractionMediaContent

Base class for media content blocks (image, audio, video, document), which carry either inline base64 Data or a Uri.

InteractionMediaDelta

Base class for media deltas (image, audio, video, document).

InteractionModalities

Known values for response modalities (response_modalities, Modality).

InteractionModalityTokens

Token count for one modality.

InteractionModelOutputStep

Model output (model_output).

InteractionModels

Model identifiers listed by the Interactions API Model enum (python-genai). The API also accepts other model IDs.

InteractionPlaceCitation

A place citation from Google Maps grounding (place_citation).

InteractionProcessingCallDelta

A processing call delta (processing_call).

InteractionProcessingCallStep

A processing call, e.g. media processing (processing_call).

InteractionProcessingResultDelta

A processing result delta (processing_result).

InteractionProcessingResultStep

The result of a processing call (processing_result).

InteractionQueryArguments

Query arguments shared by the Google Search, Google Maps and Retrieval calls.

InteractionResponseFormat

Requested output format (response_format), discriminated by type: text (structured output), image, audio or video. Formats this SDK does not know are kept as UnknownInteractionResponseFormat.

InteractionResultStep

Base class for tool result steps, which reference the call by CallId.

InteractionRetrievalCallDelta

A retrieval call delta (retrieval_call).

InteractionRetrievalCallStep

A retrieval call (retrieval_call).

InteractionRetrievalResultDelta

A retrieval result delta (retrieval_result).

InteractionRetrievalResultStep

The result of a retrieval call (retrieval_result).

InteractionRetrievalTool

The retrieval tool (retrieval).

InteractionReviewSnippet

A Google Maps review snippet.

InteractionRpcStatus

A google.rpc.Status error.

InteractionSafetySetting

A safety setting for an interaction.

InteractionServiceTiers

Known values for the service_tier field.

InteractionSpeechAnnotation

Speech synthesis metadata for a text segment (speech_metadata).

InteractionSseEvent

A Server-Sent Event of a streamed interaction, discriminated by event_type. Unknown event types are read as UnknownInteractionSseEvent.

InteractionStatusUpdateEvent

The interaction status changed (interaction.status_update).

InteractionStatuses

Known values of Status. The API may return values that are not listed here, so the status is exposed as a string.

InteractionStep

A step of an interaction (user input, model output, thought, tool call or tool result), discriminated by type. Unknown step types are read as UnknownInteractionStep.

InteractionStepDeltaData

The payload of a step.delta event, discriminated by type. Unknown delta types are read as UnknownInteractionStepDeltaData.

InteractionStepDeltaEvent

Incremental data for a step (step.delta).

InteractionStepDeltaMetadata

Metadata accompanying a streamed delta.

InteractionStepStartEvent

A step started (step.start).

InteractionStepStopEvent

A step finished (step.stop).

InteractionStepTypes

Known values of the type discriminator of InteractionStep.

InteractionTextAnnotationDelta

Annotations for the text of the step (text_annotation_delta).

InteractionTextContent

A text content block.

InteractionTextDelta

A text delta (text).

InteractionTextResponseFormat

Text output format, e.g. JSON with a schema (type: text).

InteractionThinkingLevels

Known values for ThinkingLevel.

InteractionThinkingSummaries

Known values for the thinking_summaries field.

InteractionThoughtSignatureDelta

A thought signature (thought_signature).

InteractionThoughtStep

A thought / reasoning step (thought).

InteractionThoughtSummaryDelta

A thought summary chunk (thought_summary).

InteractionTool

A tool the model may use during an interaction, discriminated by type. Unknown tool types are read as UnknownInteractionTool.

InteractionToolChoice

Tool choice: either a Mode string (see InteractionToolChoiceModes) or an AllowedTools configuration.

InteractionToolChoiceModes

Known values for the mode of InteractionToolChoice.

InteractionToolTypes

Known values of the type discriminator of InteractionTool.

InteractionTranscriptionConfig

Speech recognition (transcription) configuration.

InteractionUrlCitation

A URL citation (url_citation).

InteractionUrlContextArguments

Arguments of a URL context call.

InteractionUrlContextCallDelta

A URL context call delta (url_context_call).

InteractionUrlContextCallStep

A URL context call (url_context_call).

InteractionUrlContextResult

One fetched URL of a URL context result.

InteractionUrlContextResultDelta

A URL context result delta (url_context_result).

InteractionUrlContextResultStep

The result of a URL context call (url_context_result).

InteractionUrlContextTool

The URL context tool (url_context).

InteractionUsage

Token usage of an interaction.

InteractionUserInputStep

User input (user_input).

InteractionVertexAISearchConfig

Vertex AI Search configuration for the retrieval tool.

InteractionVideoConfig

Video generation configuration.

InteractionVideoContent

A video content block (input for video understanding, or generated video output).

InteractionVideoDelta

A video delta (video).

InteractionVideoResponseFormat

Video output format (type: video).

InteractionWebhookConfig

Webhook configuration for an interaction.

InteractionWordInfo

Word-level transcription information (word_info).

InteractionsSerializerContext

Source-generated JSON metadata for the Interactions API types. InteractionsClient puts it first in its resolver chain, so the Interactions types also work with Native AOT and trimming.

Interval

A time interval. Both ends are optional.

JobError

Represents an error that occurred during job execution.

JobStateConverter

JSON converter for JobState that handles both BATCH_STATE_* and JOB_STATE_* formats. Google AI uses BATCH_STATE_* which is converted to JOB_STATE_* for consistency.

LanguageAuto

Deprecated: Language auto-detection is now the default when LanguageCodes is omitted.

LanguageHints

Deprecated: Use LanguageCodes instead.

LatLng

An object that represents a latitude/longitude pair. This is expressed as a pair of doubles to represent degrees latitude and degrees longitude. Unless specified otherwise, this object must conform to the WGS84 standard. Values must be within normalized ranges.

ListAgentsResponse

Response of GET /agents.

ListBatchJobsResponse

Response for the list batch jobs method. Contains a paginated list of batch jobs.

ListCachedContentsResponse

Response to listing cached contents.

ListChunksResponse

Response from chunks.list containing a paginated list of Chunk. The Chunks are sorted by ascending chunk.create_time.

ListCorporaResponse

Represents a list corpora response.

ListCredentialsResponse

Response of GET /credentials.

ListDocumentsResponse

Response from documents.list containing a paginated list of Document. The Documents are sorted by ascending document.create_time.

ListEnvironmentFilesResponse

Response of GET /environments/{environment}/files/{path} (and of a file upload).

ListEnvironmentsResponse

Response of GET /environments.

ListFileSearchDocumentsResponse

Response of fileSearchStores.documents.list.

ListFileSearchStoresResponse

Response of fileSearchStores.list.

ListFilesResponse

Response to list files.

ListModelsResponse

Response to list models.

ListPermissionsResponse

Response from ListPermissions containing a paginated list of permissions. See Official API Documentation

ListTriggerExecutionsResponse

Response of GET /triggers/{id}/executions.

ListTriggersResponse

Response of GET /triggers.

ListTunedModelsResponse

Response from tunedModels.list containing a paginated list of Models.

ListTuningJobsResponse

Response for the list tuning jobs method. Contains a paginated list of tuning jobs.

ListVoicesOptions

Filters of GET /voices. Each list filter matches any of its values (case-insensitive).

ListVoicesResponse

Response of GET /voices.

LiveClientContent

Incremental update of the current conversation delivered from the client. All of the content here is unconditionally appended to the conversation history and used as part of the prompt to the model to generate content. A message here will interrupt any current model generation.

LiveClientContextUpdate

Updates to the context of the current live session. Only fields that are set will be updated. Updates are guaranteed to be processed in order with the rest of the inputs.

LiveClientMessage

Represents the payload for a bidirectional client interaction. This class is used to encapsulate data such as the setup, client content, real-time input, and tool responses that are involved in bidirectional communication.

LiveClientRealtimeInput

User input that is sent in real time. This is different from LiveClientContent in a few ways:

  • Can be sent continuously without interruption to model generation.
  • If there is a need to mix data interleaved across the LiveClientContent and the LiveClientRealtimeInput, the server attempts to optimize for best response, but there are no guarantees.
  • End of turn is not explicitly specified, but is rather derived from user activity (for example, end of speech).
  • Even before the end of turn, the data is processed incrementally to optimize for a fast start of the response from the model.
  • Is always direct user input that is sent in real time. Can be sent continuously without interruptions. The model automatically detects the beginning and the end of user speech and starts or terminates streaming the response accordingly. Data is processed incrementally as it arrives, minimizing latency.
LiveClientSetup

Message to be sent in the first and only first client message. Contains configuration that will apply for the duration of the streaming session. Clients should wait for a LiveServerSetupComplete message before sending any additional messages.

LiveClientToolResponse

Client generated response to a LiveServerToolCall received from the server. Individual FunctionResponse objects are matched to the respective FunctionCall objects by the id field. Note that in the unary and server-streaming GenerateContent APIs function calling happens by exchanging the Content parts, while in the bidi GenerateContent APIs function calling happens over these dedicated set of messages.

LiveMusicClientContent

User input to start or steer the music.

LiveMusicClientMessage

Messages sent by the client in the LiveMusicClientMessage call.

LiveMusicClientSetup

Message to be sent by the system when connecting to the API.

LiveMusicConnectParameters

Parameters for connecting to the live API.

LiveMusicFilteredPrompt

A prompt that was filtered with the reason.

LiveMusicGenerationConfig

Configuration for music generation.

LiveMusicServerContent

Server update generated by the model in response to client messages. Content is generated as quickly as possible, and not in real time. Clients may choose to buffer and play it out in real time.

LiveMusicServerMessage

Response message for the LiveMusicClientMessage call.

LiveMusicServerSetupComplete

Sent in response to a LiveMusicClientSetup message from the client.

LiveMusicSetConfigParameters

Parameters for setting config for the live music API.

LiveMusicSetWeightedPromptsParameters

Parameters for setting weighted prompts for the live music API.

LiveMusicSourceMetadata

Prompts and config used for generating this audio chunk.

LiveServerContent

Incremental server update generated by the model in response to client messages. Content is generated as quickly as possible, and not in real time. Clients may choose to buffer and play it out in real time.

LiveServerGoAway

Server notice that the connection will be closed soon. The client should reconnect (for example using a session resumption handle) before TimeLeft elapses.

LiveServerMessage

Represents a bidirectional response payload containing various types of server responses in a multimodal live session.

LiveServerSessionResumptionUpdate

Update of the session resumption state. Only sent if SessionResumptionConfig was set in the setup message.

LiveServerSetupComplete

Sent in response to a LiveClientSetup message from the client.

LiveServerToolCall

Request for the client to execute the FunctionCall and return the responses with the matching ids.

LiveServerToolCallCancellation

Notification for the client that a previously issued LiveServerToolCall with the specified ids should have been not executed and should be cancelled. If there were side-effects to those tool calls, clients may attempt to undo the tool calls. This message occurs only in cases where the clients interrupt server turns.

LogprobsCandidate

Candidate for the logprobs token and score.

LogprobsResult

Logprobs Result.

ManualRoutingMode

When manual routing is set, the specified model will be used directly.

McpServer

A remote MCP server that the model calls server-side. For client-side MCP tools see GenerativeAI.Tools.Mcp.McpTool.

MetadataFilter

User provided filter to limit retrieval based on Chunk or Document level metadata values. Example (genre = drama OR genre = action): key = "document.custom_metadata.genre" conditions = [{stringValue = "drama", operation = EQUAL}, {stringValue = "action", operation = EQUAL}]

Metric

The metric used for evaluation during model tuning.

ModalityTokenCount

Represents token counting info for a single modality.

Model

Information about a Generative Language Model.

ModelArmorConfig

Model Armor configuration for sanitizing prompts and responses. Vertex AI only.

ModelSelectionConfig

Config for model selection.

ModelStatus

Lifecycle status of the model that served a request, including deprecation and retirement information.

MultiSpeakerVoiceConfig

The configuration for the multi-speaker setup.

NoulCriteria

Descriptions of the outcomes of a NoulDecisionQuestion.

NoulDecisionAnswer

The answer to a NoulDecisionQuestion (noul).

NoulDecisionQuestion

A yes/no ("NOUL") question (noul), answered with a probability.

OAuth2CredentialConfig

OAuth 2.0 refresh-token credentials (oauth2).

OutputConfig

Config for evaluation output during model tuning.

OutputOptions

Represents the output options for the generated image.

Part

A datatype containing media that is part of a multi-part Content message. A Part consists of data which has an associated datatype.
A Part must have a fixed IANA MIME type identifying the type and subtype of the media if the InlineData field is filled with raw bytes.

PartMediaResolution

Media resolution for a single part. Overrides MediaResolution for that part.

PartialArg

A streamed piece of a function call's arguments. Vertex AI only; enabled with StreamFunctionCallArguments.

PartnerModelTuningSpec

Tuning specification for open-source and third-party Partner models.

Permission

Permission resource grants user, group or the rest of the world access to the PaLM API resource (e.g. a tuned model, corpus).

A role is a collection of permitted operations that allows users to perform specific actions on PaLM API resources. To make them available to users, groups, or service accounts, you assign roles. When you assign a role, you grant permissions that the role contains.

There are three concentric roles. Each role is a superset of the previous role's permitted operations:

  • reader can use the resource (TunedModel, Corpus) for inference
  • writer has reader's permissions and additionally can edit and share
  • owner has writer's permissions and additionally can delete
See Official API Documentation
PingWebhookRequest

Request body of POST /webhooks/{id}:ping. Currently has no fields.

PreTunedModel

A pre-tuned model for continuous tuning. Used as a base for further fine-tuning iterations.

PrebuiltVoiceConfig

The configuration for the prebuilt speaker to use.

PreferenceOptimizationHyperParameters

Hyperparameters for preference optimization.

PreferenceOptimizationSpec

Tuning spec for preference optimization.

ProactivityConfig

Configures the proactivity of the model.

PromptFeedback

A set of the feedback metadata the prompt specified in Contents.

PromptedVoice

Parameters of a voice designed from a text prompt.

QueryCorpusRequest

Represents a request to query a corpus.

QueryCorpusResponse

Represents the response from querying a corpus.

QueryDocumentRequest

Request body for performing semantic search over a Document.

QueryDocumentResponse

Response from documents.query containing a list of relevant chunks.

RagChunk

A RagChunk includes the content of a chunk of a RagFile, and associated metadata.

RagChunkPageSpan

Represents where the chunk starts and ends in the document.

RealtimeInputConfig

Configures the realtime input behavior in RealtimeInputConfig.

RegisterFilesRequest

Request body of files:register, which registers Google Cloud Storage files with the Gemini API Files service.

RegisterFilesResponse

Response of files:register.

ReinforcementTuningAutoraterScorer

Scores parsed responses with an autorater model.

ReinforcementTuningAutoraterScorerExactMatchScorer

Scores autorater responses by exact string match.

ReinforcementTuningAutoraterScorerParsedResponseConversionScorer

Scores autorater responses by converting the parsed response to a float reward clipped to [-1, 1]. This message has no fields; set an instance to select this scorer.

ReinforcementTuningCloudRunRewardScorer

Scores parsed responses by calling a Cloud Run service.

ReinforcementTuningCodeExecutionRewardScorer

Scores parsed responses by running a Python code snippet.

ReinforcementTuningExample

An example (prompt) used to validate a reinforcement tuning reward configuration.

ReinforcementTuningHyperParameters

Hyperparameters for reinforcement tuning.

ReinforcementTuningParseResponseConfig

Defines how a response is parsed before scoring.

ReinforcementTuningRewardInfo

The reward info for a single reward function.

ReinforcementTuningSpec

Tuning spec for reinforcement tuning (Vertex AI only).

ReinforcementTuningStringMatchRewardScorer

Scores parsed responses by simple string matching.

ReinforcementTuningStringMatchRewardScorerJsonMatchExpression

Parses the response as JSON, finds the first-level value for KeyName and evaluates it.

ReinforcementTuningStringMatchRewardScorerStringMatchExpression

Evaluates a parsed response using a match operation against an expression.

RelevantChunk

The information for a chunk relevant to a query.

RemoteFile

A file uploaded to the API.

ReplicatedVoice

Parameters of a voice replicated from recorded audio.

ReplicatedVoiceConfig

Configuration for speaking with a voice replicated from an audio sample.

RequestsWrapper

Wrapper for requests array (Google AI nested structure).

ResponseFormat

Output format for one response modality. Set only the member for the modality it applies to. Vertex AI only.

RetrievalConfig

Retrieval config for tool configuration.

RetrievalMetadata

Metadata related to retrieval in the grounding flow.

RotateSigningSecretRequest

Request body of POST /webhooks/{id}:rotateSigningSecret.

RoutingConfig

The configuration for routing the request to a specific model.

SafetyAttributes

Represents the safety attributes of a generated image.

SafetyRating

Safety rating for a piece of content. The safety rating contains the category of harm and the harm probability level in that category for a piece of content. Content is classified for safety across a number of harm categories and the probability of the harm classification is included here.

SafetySetting

Safety setting, affecting the safety-blocking behavior. Passing a safety setting for a category changes the allowed probability that content is blocked.

Schema

The Schema object allows the definition of input and output data types. These types can be objects, but also primitives and arrays. Represents a select subset of an OpenAPI 3.0 schema object.

SchemaSourceGenerationContext

The SchemaSourceGenerationContext class is a custom source generation context for improving the performance of JSON serialization and deserialization. This is achieved by leveraging the JsonSourceGenerationOptionsAttribute to configure source generation options and define types for serialization at compile-time.

ScoreDecisionAnswer

The answer to a ScoreDecisionQuestion (score).

ScoreDecisionQuestion

A scoring question (score).

SearchEntryPoint

Google search entry point.

SearchTypes

The kinds of Google Search the model may use. Leave unset for web search only.

Segment

Segment of the content.

SemanticRetrieverChunk

Identifier for a Chunk retrieved via Semantic Retriever specified in the GenerateAnswerRequest using SemanticRetrieverConfig.

SemanticRetrieverConfig

Configuration for retrieving grounding content from a Corpus or Document created using the Semantic Retriever API.

ServiceTiers

Values for ServiceTier.

SessionResumptionConfig

Configuration of session resumption mechanism. Included in LiveConnectConfig.session_resumption. If included server will send LiveServerSessionResumptionUpdate messages.

SigningSecret

Represents a signing secret used to verify webhook payloads.

SingleEmbedContentResponse

Response for a single embedding content request.

SingleReinforcementTuningRewardConfig

A single reward function configuration for reinforcement tuning.

SlidingWindow

Context window will be truncated by keeping only suffix of it. Context window will always be cut at start of USER role turn. System instructions and LiveClientSetup.prefix_turns will not be subject to the sliding window mechanism, they will always stay at the beginning of context window.

SpeakerVoiceConfig

The configuration for a single speaker in a multi speaker setup.

SpeechConfig

The speech generation config.

SpeechMetadata

Speaker and style metadata for speech generated in a part.

Status

The Status type defines a logical error model that is suitable for different programming environments, including REST APIs and RPC APIs. It is used by gRPC. Each Status message contains three pieces of data: error code, error message, and error details. You can find out more about this error model and how to work with it in the API Design Guide.

StreamableHttpTransport

Streamable HTTP transport settings for a server-side MCP server.

StringList

User provided string values assigned to a single metadata key.

SupervisedHyperParameters

Hyperparameters for Supervised Fine-Tuning (SFT).

SupervisedTuningDataStats

Tuning data statistics for Supervised Fine-Tuning. Provides comprehensive analytics about the training dataset.

SupervisedTuningDatasetDistribution

Dataset distribution for Supervised Tuning. Provides statistical analysis of the training dataset.

SupervisedTuningDatasetDistributionDatasetBucket

Dataset bucket used to create a histogram for the distribution given a population of values.

SupervisedTuningSpec

Tuning specification for Supervised Fine-Tuning for first-party models.

TextResponseFormat

Format of text output.

ThinkingConfig

Config for thinking features.

Timestamp

A Timestamp represents a point in time independent of any time zone or calendar, represented as seconds and fractions of seconds at nanosecond resolution in UTC Epoch time.

TimestampJsonConverter

JSON converter for Timestamp.

TokensInfo

Tokens info with a list of tokens and the corresponding list of token ids.

Tool

Tool details that the model may use to generate response. A Tool is a piece of code that enables the system to interact with external systems to perform an action, or set of actions, outside of knowledge and scope of the model.

ToolCall

A call to a server-side tool (for example Google Search or File Search) made by the model. Returned when IncludeServerSideToolInvocations is true.

ToolConfig

The Tool configuration containing parameters for specifying Tool use in the request.

ToolExaAiSearch

Exa AI web search tool. Vertex AI only.

ToolParallelAiSearch

Parallel AI web search tool. Vertex AI only.

ToolResponse

The result of a server-side tool call. Returned when IncludeServerSideToolInvocations is true.

TopCandidates

Candidates with top log probabilities at each decoding step.

Transcription

Audio transcription in Server Content.

TranslationConfig

Config for stream translation. Sent on the wire as setup.generationConfig.translationConfig.

Trigger

A trigger (/triggers): runs an agent interaction on a cron schedule.

TriggerExecution

A run of a Trigger.

TriggerExecutionStatuses

Known trigger execution statuses.

TriggerStatuses

Known trigger statuses.

TunedModel

TunedModel for the Tuned Model of a Tuning Job. Represents the output model produced by a tuning job.

TunedModelCheckpoint

TunedModelCheckpoint for the Tuned Model of a Tuning Job. Represents a specific checkpoint during the tuning process.

TunedModelInfo

Information about the tuned model from the base model.

TunedModelSource

Tuned model as a source for training a new model.

TuningDataStats

The tuning data statistics associated with the tuning job.

TuningDataset

The training dataset of a tuning job.

TuningExample

A single example for tuning.

TuningExamples

A set of tuning examples. Can be training or validation data.

TuningJob

Represents a model tuning job. A tuning job is used to fine-tune a base model with custom training data to create a specialized model tailored to specific tasks or domains.

TuningJobMetadata

Tuning job metadata.

TuningSerializerContext

Source-generated serialization metadata for the tuning types used by TuningClient.

TuningSnapshot

Record for a single tuning step.

TuningTask

Tuning tasks that create tuned models.

TuningValidationDataset

The validation dataset of a tuning job (Vertex AI only).

TypesSerializerContext

JSON serializer context for GenerativeAI types, providing source-generated serialization metadata.

UnknownCredentialConfig

A credential configuration whose type this SDK does not know; the raw JSON is kept and sent as is.

UnknownDecisionAnswer

An answer whose type this SDK does not know; the raw JSON is kept.

UnknownDecisionQuestion

A question whose type this SDK does not know; the raw JSON is kept.

UnknownInteractionAgentConfig

An agent configuration whose type this SDK does not know.

UnknownInteractionAnnotation

An annotation whose type this SDK does not know.

UnknownInteractionContent

A content block whose type this SDK does not know. Raw keeps the original JSON and is written back unchanged.

UnknownInteractionResponseFormat

A response format this SDK does not know, or a free-form JSON object.

UnknownInteractionSseEvent

An event whose event_type this SDK does not know.

UnknownInteractionStep

A step whose type this SDK does not know. Raw keeps the original JSON and is written back unchanged (for example when the step is passed back as input).

UnknownInteractionStepDeltaData

A delta whose type this SDK does not know.

UnknownInteractionTool

A tool whose type this SDK does not know.

UpdateChunkRequest

Represents a request to update a chunk.

UpdateModelConfig

The fields to update on a (tuned) model, mirroring UpdateModelConfig of the google-genai SDK. Only the fields that are set are sent.

UpdateTriggerRequest

Request body of PATCH /triggers/{id}.

UploadFileInformation

Represents information about a file to be uploaded.

UploadFileRequest

Represents a request to upload a local file in order to create a File resource.

UploadFileResponse

Represents the response containing details about a file uploaded through the File API.

UploadToFileSearchStoreConfig

Optional settings for uploading a file directly into a FileSearchStore. Sent as the JSON metadata of the uploadToFileSearchStore request.

UpscaleConfig

Represents the Upscale configuration.

UrlContext

Tool to support URL context retrieval.

UrlContextMetadata

Metadata related to url context retrieval tool.

UrlMetadata

Context for a single url retrieval.

UsageMetadata

Metadata on the generation request's token usage.

ValidateRewardRequest

Request body of tuningJobs:validateReinforcementTuningReward.

ValidateRewardResponse

Response of tuningJobs:validateReinforcementTuningReward.

VeoHyperParameters

Hyperparameters for Veo model tuning.

VeoLoraTuningSpec

Tuning spec for Veo LoRA tuning.

VeoTuningSpec

Tuning specification for Veo video generation model tuning.

VertexAIListBatchJobsResponse

Response for the list batch jobs method (Vertex AI wire format).

VertexBatchEmbedContentsRequest

Request body for the Vertex AI batchEmbedContents method.

VertexBatchRequest

Request for batch job creation (Vertex AI format).

VertexEmbedContentConfig

Embedding options sent in the Vertex AI embedContent request format. Built from EmbedContentRequest by the SDK; you normally do not create this directly.

VertexEmbedContentRequest

Request body for the Vertex AI embedContent method, used by Gemini embedding models on Vertex AI.

VertexEmbedContentResponse

Response of the Vertex AI embedContent method.

VertexEmbeddingInstance

One text to embed with the Vertex AI predict method.

VertexEmbeddingParameters

Options for the Vertex AI embedding predict method.

VertexEmbeddingPredictRequest

Request body for the Vertex AI predict method, used by text embedding models such as text-embedding-005 and textembedding-gecko.

VertexEmbeddingPredictResponse

Response of the Vertex AI embedding predict method.

VertexEmbeddingPrediction

One prediction returned by the Vertex AI embedding predict method.

VertexGenerateVideosPayload

Internal representation of the payload for Vertex AI video generation requests.

VertexInputConfig

Input configuration for Vertex AI batch jobs.

VertexOutputConfig

Output configuration for Vertex AI batch jobs.

VertexPredictEmbedding

An embedding returned by the Vertex AI embedding predict method.

VertexPredictEmbeddingStatistics

Input statistics returned by the Vertex AI embedding predict method (snake_case on the wire).

VertexRetrievalTool

Represents a retrieval tool that can access data sources powered by Vertex AI Search or Vertex RAG store.

Video

Represents a video resource, which can be referenced via a URI or contained as raw byte data.

VideoGenerationMask

A mask for video generation.

VideoGenerationReferenceImage

A reference image for video generation.

VideoInstance

Internal representation of a single instance within the Vertex AI payload.

VideoMetadata

Metadata for a video File.

VideoMetadataResult

Metadata describes the input video content. Describes how the video in the Part should be used by the model.

VideoMetadataVideoGeneration

Metadata describes the input video content.

VideoParameters

Internal representation of the parameters within the Vertex AI payload. Mirrors GenerateVideosConfig but ensures correct JsonPropertyNames for serialization.

VideoResponseFormat

Format of video output.

VisionGenerativeModelResult

Represents the result of a vision generative model.

Voice

A voice: a custom voice created with POST /voices (designed from a prompt or replicated from audio) or a prebuilt system voice from Google's voice catalog (returned by GET /voices). Use its Key / Id as the voice of a text-to-speech request.

VoiceActivity

Voice activity signal sent by the server.

VoiceActivityDetectionSignal

Voice activity detection (VAD) signal sent by the server. Allowlisted only.

VoiceAudioData

Inline audio data.

VoiceConfig

The configuration for the voice to use.

VoiceConsentSignature

Proof that voice-replication consent was verified.

VoicePitches

Known voice pitches.

VoiceTypes

Known voice types.

VqaImage

Represents the image data for the visual question answering request.

VqaInstance

Represents a single instance in the visual question answering request.

VqaParameters

Represents the parameters for the visual question answering request.

VqaRequest

Represents the request body for the visual question answering API.

VqaResponse

Represents the response body for the visual question answering API.

Web

Chunk from the web.

WebSearch

Web search settings. Currently has no options; set an instance to enable web search.

Webhook

A Webhook resource: an endpoint that receives event notifications (for example when a batch job or a background interaction finishes). The Webhooks API uses snake_case field names on the wire.

WebhookConfig

Per-request webhook configuration for long-running operations (e.g. batch jobs): notifications are sent when the operation completes. Only supported by the Gemini Developer API.

WebhookEvents

Known event types a Webhook can subscribe to (see SubscribedEvents). The API may add new events, so events are exposed as strings.

WebhookListResponse

Response of GET /webhooks.

WebhookPingResponse

Response of POST /webhooks/{id}:ping. Currently has no fields.

WebhookRevocationBehaviors

Known values of RevocationBehavior.

WebhookRotateSigningSecretResponse

Response of POST /webhooks/{id}:rotateSigningSecret.

WebhookStates

Known values of State and State. The API may return values that are not listed here, so the state is exposed as a string.

WebhookUpdate

The request body of PATCH /webhooks/{id} (python-genai WebhookUpdate). Only the fields that are set are sent; combine with an update mask to control which fields are replaced.

WebhooksSerializerContext

Source-generated JSON metadata for the Webhooks API types. WebhooksClient puts it first in its resolver chain, so the Webhooks types also work with Native AOT and trimming.

WeightedPrompt

Maps a prompt to a relative weight to steer music generation.

WhiteSpaceConfig

Configuration for the white space chunking algorithm.

WordInfo

Information about a single recognized word in a Transcription.

Interfaces

IContentsRequest

Represents a request containing a collection of contents.

Enums

ActivityHandling

The different ways of handling user activity.

AdapterSize

Optional. Adapter size for tuning.

AnswerStyle

Style for grounded answers.

ApiSpec

The API spec that the external API implements.

AspectRatio

Aspect ratio of generated images or video.

AudioTranscriptionConfigMode

Transcription mode used by Mode.

AuthType

Type of auth scheme.

Behavior

Defines the function behavior. Defaults to BLOCKING.

BlockReason

Specifies the reason why the prompt was blocked.

ChunkState

States for the lifecycle of a Chunk.

Delivery

How generated media is returned.

DynamicRetrievalMode

The mode of the predictor to be used in dynamic retrieval.

EndSensitivity

End of speech sensitivity.

Environment

The environment being operated.

FeatureSelectionPreference

Options for feature selection preference.

FileSearchDocumentState

Processing state of a FileSearchDocument.

FileState

States for the lifecycle of a File.

FinishReason

Defines the reason why the model stopped generating tokens.

FunctionCallingMode

Defines the execution behavior for function calling by defining the execution mode.

FunctionResponseScheduling

Specifies how the response should be scheduled in the conversation.

GranteeType

Defines types of the grantee of this permission. See Official API Documentation

HarmBlockMethod

Specifies if the threshold is used for probability or severity score. If not specified, the threshold is used for probability score.

HarmBlockThreshold

Block at and beyond a specified harm probability.

HarmCategory

The category of a rating. These categories cover various kinds of harms that developers may wish to adjust.

HarmProbability

The probability that a piece of content is harmful. The classification system gives the probability of the content being unsafe. This does not indicate the severity of harm for a piece of content.

HarmSeverity

Output only. Harm severity levels in the content.

ImageSafetySetting

Represents the safety filter level.

ImageSize

Size of generated images.

InteractionStatus

The different activity states of the live session.

JobState

Job state enumeration representing the lifecycle states of a job. Supports both BATCH_STATE_* (Google AI) and JOB_STATE_* (Vertex AI) formats.

Language

Supported programming languages for the generated code.

LiveMusicPlaybackControl

The playback control signal to apply to the music generation.

MatchOperation

The match operation used by a reinforcement tuning string match expression.

MediaProcessing

How the model processes the media in a part.

MediaResolution

Media resolution for the input media.

Modality

Content Part modality.

ModelRoutingPreference

Model routing preference options.

ModelStage

Lifecycle stage of a model.

MusicGenerationMode

The mode of music generation.

Operator

Defines the valid operators that can be applied to a key-value pair.

Outcome

Enumeration of possible outcomes of the code execution.

PartMediaResolutionLevel

Resolution level for the media in a single part.

PersonGeneration

Represents the allowed generation of people by the model.

PhishBlockThreshold

Confidence threshold for blocking phishing sites in search results. Vertex AI only.

ProminentPeople

Whether images of prominent people may be generated. Vertex AI only.

ReinforcementTuningThinkingLevel

The maximum thinking depth used during reinforcement tuning.

ResponseParseType

Defines how a sample response is parsed before scoring in reinforcement tuning.

Role

Defines the role granted by this permission. See Official API Documentation

SafetyPolicy

Computer Use safety policies that can be disabled through DisabledSafetyPolicies.

Scale

Scale of the generated music.

SchemaType

Type contains the list of OpenAPI data types as defined by https://spec.openapis.org/oas/v3.0.3#data-types.

SessionResumptionStatus

Represents the status of a session resumption attempt.

Source

Source of the File.

StartSensitivity

Start of speech sensitivity.

TaskType

Specifies the type of task for which the embedding will be used.

ThinkingLevel

Controls the depth of thinking/reasoning for models that support thinking features. Used with Gemini 3 and later models.

ToolType

Type of a server-side tool.

TrafficType

Output only. Traffic type. This shows whether a request consumes Pay-As-You-Go or Provisioned Throughput quota.

TuningJobState

The detailed state of a tuning job while the overall JobState is running.

TuningMethod

The tuning method used by a tuning job.

TuningMode

Tuning mode enumeration.

TuningSpeed

The speed of a Veo tuning job.

TuningState

The state of the tuned model.

TurnCompleteReason

The reason why the turn is complete.

TurnCoverage

Options about which input is included in the user's turn.

UrlRetrievalStatus

Status of the url retrieval.

VadSignalType

The type of the VAD signal.

VeoTuningTask

The tuning task of a Veo tuning job.

VideoAspectRatio

Defines the possible aspect ratios for generated videos.

VideoCompressionQuality

Enum that controls the compression quality of the generated videos.

VideoGenerationMaskMode

Enum for the mask mode of a video generation mask.

VideoGenerationReferenceType

Enum for the reference type of a video generation reference image.

VideoOrientation

The orientation of the videos produced by a Veo LoRA tuned model.

VideoPersonGeneration

Allows the generation of all types of people, including unrestricted content, in videos.

VideoResolution

Defines the possible resolutions for generated videos. Actual available values depend on the model.

VoiceActivityType

The type of the voice activity signal.