The model’s response to a generate content request.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructGenerateContentResponse:Sendable

extensionGenerateContentResponse:Decodable

A struct representing a possible reply to a content generation prompt. Each content generationprompt may produce multiple candidate responses.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructCandidate:Sendable

extensionCandidate:Decodable

A collection of source attributions for a piece of content.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructCitationMetadata:Sendable

extensionCitationMetadata:Decodable

A struct describing a source attribution.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructCitation:Sendable,Equatable

extensionCitation:Decodable

A value enumerating possible reasons for a model to terminate a content generation request.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructFinishReason:DecodableProtoEnum,Hashable,Sendable

A metadata struct containing any feedback the model had on the prompt it was provided.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructPromptFeedback:Sendable

extensionPromptFeedback:Decodable

Metadata returned to the client when grounding is enabled.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructGroundingMetadata:Sendable,Equatable,Hashable

extensionGroundingMetadata:Decodable

Represents a specific segment within aModelContent struct, often used to pinpoint theexact location of text or data that grounding information refers to.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructSegment:Sendable,Equatable,Hashable

extensionSegment:Decodable

A struct defining model parameters to be used when sending generative AIrequests to the backend model.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructGenerationConfig:Sendable

extensionGenerationConfig:Encodable

Configuration parameters for sending requests to the backend.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructRequestOptions:Sendable

extensionRequestOptions:Equatable

Represents token counting info for a single modality.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructModalityTokenCount:Sendable

extensionModalityTokenCount:Decodable

Content part modality.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructContentModality:DecodableProtoEnum,Hashable,Sendable

A type describing data in media formats interpretable by an AI model. Each generative AIrequest or response contains anArray ofModelContents, and eachModelContent valuemay comprise multiple heterogeneousParts.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructModelContent:Equatable,Sendable

extensionModelContent:Codable

A type defining potentially harmful media categories and their model-assigned ratings. A valueof this type may be assigned to a category for every model-generated response, not justresponses that exceed a certain threshold.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructSafetyRating:Equatable,Hashable,Sendable

extensionSafetyRating:Decodable

A type used to specify a threshold for harmful content, beyond which the model will return afallback response instead of generated content.

Seesafety settings for Geminimodels formore details.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructSafetySetting:Sendable

extensionSafetySetting:Encodable

Categories describing the potential harm a piece of content may pose.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructHarmCategory:CodableProtoEnum,Hashable,Sendable

Structured representation of a function declaration.

ThisFunctionDeclaration is a representation of a block of code that can be used as aToolby the model and executed by the client.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructFunctionDeclaration:Sendable

extensionFunctionDeclaration:Encodable

A tool that allows the generative model to connect to Google Search to access and incorporateup-to-date information from the web into its responses.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructGoogleSearch:Sendable

extensionGoogleSearch:Encodable

A helper tool that the model may use when generating responses.

ATool is a piece of code that enables the system to interact with external systems to performan action, or set of actions, outside of knowledge and scope of the model.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructTool:Sendable

extensionTool:Encodable

Configuration for specifying function calling behavior.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructFunctionCallingConfig:Sendable

extensionFunctionCallingConfig:Encodable

Tool configuration for anyTool specified in the request.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructToolConfig:Sendable

extensionToolConfig:Encodable

The model’s response to a count tokens request.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructCountTokensResponse:Sendable

extensionCountTokensResponse:Decodable

Represents available backend APIs for the Firebase AI SDK.

Declaration

publicstructBackend

An aspect ratio for images generated by Imagen.

To specify an aspect ratio for generated images, setaspectRatio inyourImagenGenerationConfig. See theClouddocumentationfor more details and examples of the supported aspect ratios.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructImagenAspectRatio:Sendable

Configuration options for generating images with Imagen.

SeeParameters for Imagenmodels tolearn about parameters available for use with Imagen models, including how to configure them.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructImagenGenerationConfig

A response from a request to generate images with Imagen.

The type placeholderT is an image type; this is currently always anImagenInlineImage.

This type is returned from:

• generateImages(prompt:) whereT isImagenInlineImage

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructImagenGenerationResponse<T>:SendablewhereT:Sendable

extensionImagenGenerationResponse:DecodablewhereT:Decodable

An image format for images generated by Imagen.

To specify an image format for generated images, setimageFormat inyourImagenGenerationConfig. See theClouddocumentationfor more details.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructImagenImageFormat

An error that occurs when image generation fails due to all generated images being blocked.

The images may have been blocked due to the specifiedImagenSafetyFilterLevel, theImagenPersonFilterLevel, or filtering included in the model. These filter levels may beadjusted in yourImagenSafetySettings. See theResponsible AI and usage guidelines forImagenfor more details.

Declaration

publicstructImagenImagesBlockedError:Error

extensionImagenImagesBlockedError:CustomNSError

An image generated by Imagen, represented as inline data.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructImagenInlineImage:Sendable

extensionImagenInlineImage:Equatable

extensionImagenInlineImage:Decodable

A filter level controlling whether generation of images containing people or faces is allowed.

See thepersonGenerationdocumentation for more details.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructImagenPersonFilterLevel:ProtoEnum

A filter level controlling how aggressively to filter sensitive content.

Text prompts provided as inputs and images (generated or uploaded) through Imagen on Vertex AIare assessed against a list of safety filters, which include ‘harmful categories’ (for example,violence,sexual,derogatory, andtoxic). This filter level controls how aggressively tofilter out potentially harmful content from responses. See thesafetySettingdocumentation and theResponsible AI and usageguidelinesfor more details.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructImagenSafetyFilterLevel:ProtoEnum,Sendable

Settings for controlling the aggressiveness of filtering out sensitive content.

See theResponsible AI and usageguidelinesfor more details.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructImagenSafetySettings

Configuration options for audio transcriptions when communicating with a model that supports theGemini Live API.

While there are not currently any options, this will likely change in the future. For now, justproviding an instance of this struct will enable audio transcriptions for the correspondinginput or output fields.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructAudioTranscriptionConfig:Sendable

Text transcription of some audio form during a live interaction with the model.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructLiveAudioTranscription:Sendable

Configuration options for live content generation.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructLiveGenerationConfig:Sendable

Incremental server update generated by the model in response to clientmessages.

Content is generated as quickly as possible, and not in realtime. Clientsmay choose to buffer and play it out in realtime.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructLiveServerContent:Sendable

Server will not be able to service client soon.

To learn more about session limits, see the docs onMaximum session duration.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructLiveServerGoingAwayNotice:Sendable

Update from the server, generated from the model in response to client messages.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructLiveServerMessage:Sendable

Request for the client to execute the providedfunctionCalls.

The client should return matchingFunctionResponsePart, where thefunctionId fields correspond to individualFunctionCallParts.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)@available(watchOS,unavailable)publicstructLiveServerToolCall:Sendable

Notification for the client to cancel a previous function call fromLiveServerToolCall.

The client does not need to sendFunctionResponseParts for the cancelledFunctionCallParts.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructLiveServerToolCallCancellation:Sendable

The model sent a message that the SDK failed to parse.

This may indicate that the SDK version needs updating, a model is too old for the current SDKversion, or that the model is justnot supported.

Check theNSUnderlyingErrorKey entry inerrorUserInfofor the error that caused this.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructLiveSessionUnsupportedMessageError:Error,Sendable,CustomNSError

The live session was closed, because the network connection was lost.

Check theNSUnderlyingErrorKey entry inerrorUserInfo forthe error that caused this.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructLiveSessionLostConnectionError:Error,Sendable,CustomNSError

The live session was closed, but not for a reason the SDK expected.

Check theNSUnderlyingErrorKey entry inerrorUserInfofor the error that caused this.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructLiveSessionUnexpectedClosureError:Error,Sendable,CustomNSError

The model refused our request to setup a live session.

This can occur due to the model not supporting the requested response modalities, the projectnot having access to the model, the model being invalid, or some internal error.

Check theNSUnderlyingErrorKey entry inerrorUserInfo for the errorthat caused this.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructLiveSessionSetupError:Error,Sendable,CustomNSError

Configuration for controlling the voice of the model during conversation.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,*)@available(watchOS,unavailable)publicstructSpeechConfig:Sendable

A text part containing a string value.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructTextPart:Part

A data part that is provided inline in requests.

Data provided as an inline data part is encoded as base64 and included directly (inline) in therequest. For large files, seeFileDataPart which references content by URI instead ofincluding the data in the request.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructInlineDataPart:Part

File data stored in Cloud Storage for Firebase, referenced by URI.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructFileDataPart:Part

A predicted function call returned from the model.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructFunctionCallPart:Part

Result output from a function call.

Contains a string representing theFunctionDeclaration.name and a structured JSON objectcontaining any output from the function is used as context to the model. This should contain theresult of aFunctionCallPart made based on model prediction.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructFunctionResponsePart:Part

A part containing code that was executed by the model.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructExecutableCodePart:Part

The result of executing code.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructCodeExecutionResultPart:Part

Represents the different types, or modalities, of data that a model can produce as output.

To configure the desired output modalities for model requests, set theresponseModalitiesparameter when initializing aGenerationConfig. See themultimodalresponsesdocumentation for more details.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructResponseModality:EncodableProtoEnum,Sendable

Configuration for controlling the “thinking” behavior of compatible Gemini models.

Certain models, like Gemini 2.5 Flash and Pro, utilize a thinking process before generating aresponse. This allows them to reason through complex problems and plan a more coherent andaccurate answer.

Declaration

publicstructThinkingConfig:Sendable

extensionThinkingConfig:Encodable

A tool that allows the model to execute code.

This tool can be used to solve complex problems, for example, by generating and executing Pythoncode to solve a math problem.

Declaration

publicstructCodeExecution:Sendable,Encodable

Metadata related to theurlContext() tool.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructURLContextMetadata:Sendable,Hashable

extensionURLContextMetadata:Decodable

Metadata for a single URL retrieved by theurlContext() tool.

Declaration

@available(iOS15.0,macOS12.0,tvOS15.0,watchOS8.0,*)publicstructURLMetadata:Sendable,Hashable

extensionURLMetadata:Decodable