IMPORTANT: In July 2025, we stopped releasing new versions of KTX modules, and we removed the KTX libraries from the Firebase Android BoM (v34.0.0). If you use KTX APIs from the previously released KTX modules, we recommend that you migrate your app to use KTX APIs from the main modules instead. For details, see the FAQ about this initiative.

FirebaseAI

class FirebaseAI

Entry point for all Firebase AI functionality.

Summary

Public companion functions
`FirebaseAI`	`getInstance(app: FirebaseApp)` The `FirebaseAI` instance for the provided `FirebaseApp` using the Google AI Backend.
`FirebaseAI`	`getInstance(app: FirebaseApp, backend: GenerativeBackend)` Returns the `FirebaseAI` instance for the provided `FirebaseApp` and `backend`.
`FirebaseAI`	`getInstance( app: FirebaseApp, backend: GenerativeBackend, useLimitedUseAppCheckTokens: Boolean )` Returns the `FirebaseAI` instance for the provided `FirebaseApp` and `backend`.

Public companion properties
`FirebaseAI`	`instance` The `FirebaseAI` instance for the default `FirebaseApp` using the Google AI Backend.

Public functions
`GenerativeModel`	`generativeModel( modelName: String, generationConfig: GenerationConfig?, safetySettings: List<SafetySetting>?, tools: List<Tool>?, toolConfig: ToolConfig?, systemInstruction: Content?, requestOptions: RequestOptions )` Instantiates a new `GenerativeModel` given the provided parameters.
`ImagenModel`	`@PublicPreviewAPI imagenModel( modelName: String, generationConfig: ImagenGenerationConfig?, safetySettings: ImagenSafetySettings?, requestOptions: RequestOptions )` Instantiates a new `ImagenModel` given the provided parameters.
`LiveGenerativeModel`	`@PublicPreviewAPI liveModel( modelName: String, generationConfig: LiveGenerationConfig?, tools: List<Tool>?, systemInstruction: Content?, requestOptions: RequestOptions )` Instantiates a new `LiveGenerationConfig` given the provided parameters.

Public companion functions

getInstance

fun getInstance(app: FirebaseApp): FirebaseAI

The FirebaseAI instance for the provided FirebaseApp using the Google AI Backend.

getInstance

fun getInstance(app: FirebaseApp = Firebase.app, backend: GenerativeBackend): FirebaseAI

Returns the FirebaseAI instance for the provided FirebaseApp and backend.

Parameters
`backend: GenerativeBackend`	the backend reference to make generative AI requests to.

getInstance

fun getInstance(
    app: FirebaseApp = Firebase.app,
    backend: GenerativeBackend,
    useLimitedUseAppCheckTokens: Boolean
): FirebaseAI

Returns the FirebaseAI instance for the provided FirebaseApp and backend.

Parameters

backend: GenerativeBackend

the backend reference to make generative AI requests to.

useLimitedUseAppCheckTokens: Boolean

when sending tokens to the backend, this option enables the usage of App Check's limited-use tokens instead of the standard cached tokens.

A new limited-use tokens will be generated for each request; providing a smaller attack surface for malicious parties to hijack tokens. When used alongside replay protection, limited-use tokens are also consumed after each request, ensuring they can't be used again.

This flag is set to false by default.

Important: Replay protection is not currently supported for the FirebaseAI backend. While this feature is being developed, you can still migrate to using limited-use tokens. Because limited-use tokens are backwards compatible, you can still use them without replay protection. Due to their shorter TTL over standard App Check tokens, they still provide a security benefit. Migrating to limited-use tokens sooner minimizes disruption when support for replay protection is added.

Public companion properties

instance

val instance: FirebaseAI

The FirebaseAI instance for the default FirebaseApp using the Google AI Backend.

Public functions

generativeModel

fun generativeModel(
    modelName: String,
    generationConfig: GenerationConfig? = null,
    safetySettings: List<SafetySetting>? = null,
    tools: List<Tool>? = null,
    toolConfig: ToolConfig? = null,
    systemInstruction: Content? = null,
    requestOptions: RequestOptions = RequestOptions()
): GenerativeModel

Instantiates a new GenerativeModel given the provided parameters.

Parameters
`modelName: String`	The name of the model to use. See the documentation for a list of supported models.
`generationConfig: GenerationConfig? = null`	The configuration parameters to use for content generation.
`safetySettings: List<SafetySetting>? = null`	The safety bounds the model will abide to during content generation.
`tools: List<Tool>? = null`	A list of `Tool`s the model may use to generate content.
`toolConfig: ToolConfig? = null`	The `ToolConfig` that defines how the model handles the tools provided.
`systemInstruction: Content? = null`	`Content` instructions that direct the model to behave a certain way. Currently only text content is supported.
`requestOptions: RequestOptions = RequestOptions()`	Configuration options for sending requests to the backend.

Returns
`GenerativeModel`	The initialized `GenerativeModel` instance.

imagenModel

@PublicPreviewAPI
fun imagenModel(
    modelName: String,
    generationConfig: ImagenGenerationConfig? = null,
    safetySettings: ImagenSafetySettings? = null,
    requestOptions: RequestOptions = RequestOptions()
): ImagenModel

Instantiates a new ImagenModel given the provided parameters.

Parameters
`modelName: String`	The name of the model to use. See the documentation for a list of supported models.
`generationConfig: ImagenGenerationConfig? = null`	The configuration parameters to use for image generation.
`safetySettings: ImagenSafetySettings? = null`	The safety bounds the model will abide by during image generation.
`requestOptions: RequestOptions = RequestOptions()`	Configuration options for sending requests to the backend.

Returns
`ImagenModel`	The initialized `ImagenModel` instance.

liveModel

@PublicPreviewAPI
fun liveModel(
    modelName: String,
    generationConfig: LiveGenerationConfig? = null,
    tools: List<Tool>? = null,
    systemInstruction: Content? = null,
    requestOptions: RequestOptions = RequestOptions()
): LiveGenerativeModel

Instantiates a new LiveGenerationConfig given the provided parameters.

Parameters
`modelName: String`	The name of the model to use. See the documentation for a list of supported models.
`generationConfig: LiveGenerationConfig? = null`	The configuration parameters to use for content generation.
`tools: List<Tool>? = null`	A list of `Tool`s the model may use to generate content.
`systemInstruction: Content? = null`	`Content` instructions that direct the model to behave a certain way. Currently only text content is supported.
`requestOptions: RequestOptions = RequestOptions()`	Configuration options for sending requests to the backend.

Returns
`LiveGenerativeModel`	The initialized `LiveGenerativeModel` instance.