@ai-sdk/provider 3.0.12 → 3.0.13

package/CHANGELOG.md

@@ -1,5 +1,11 @@
 # @ai-sdk/provider
 
+## 3.0.13
+
+### Patch Changes
+
+- fa850e6: feat (video): add first-class `frameImages` and `inputReferences` call options for video generation
+
 ## 3.0.12
 
 ### Patch Changes

package/dist/index.d.mts

@@ -3843,6 +3843,27 @@ type VideoModelV3File = {
     providerOptions?: SharedV3ProviderMetadata;
 };
 
+/**
+ * The role a frame image plays in video generation.
+ *
+ * - `first_frame`: the starting frame the model animates from
+ * - `last_frame`: the ending frame the model animates towards
+ */
+type VideoModelV3FrameType = 'first_frame' | 'last_frame';
+/**
+ * A role-tagged image input for image-to-video and first-last-frame generation.
+ */
+type VideoModelV3FrameImage = {
+    /**
+     * The image file used for this frame.
+     */
+    image: VideoModelV3File;
+    /**
+     * Which frame this image represents.
+     */
+    frameType: VideoModelV3FrameType;
+};
+
 type VideoModelV3CallOptions = {
     /**
      * Text prompt for the video generation.
@@ -3888,6 +3909,16 @@ type VideoModelV3CallOptions = {
      * The image serves as the starting frame that the model will animate.
      */
     image: VideoModelV3File | undefined;
+    /**
+     * Role-tagged image inputs for first-last-frame generation.
+     * Each entry declares whether it is the `first_frame` or the
+     * `last_frame` of the generated video.
+     */
+    frameImages: Array<VideoModelV3FrameImage> | undefined;
+    /**
+     * Reference image inputs for reference-to-video generation.
+     */
+    inputReferences: Array<VideoModelV3File> | undefined;
     /**
      * Whether the model should generate audio alongside the video.
      */
@@ -4031,4 +4062,4 @@ type VideoModelV3 = {
     }>;
 };
 
-export { AISDKError, APICallError, type EmbeddingModelV2, type EmbeddingModelV2Embedding, type EmbeddingModelV3, type EmbeddingModelV3CallOptions, type EmbeddingModelV3Embedding, type EmbeddingModelV3Middleware, type EmbeddingModelV3Result, EmptyResponseBodyError, type VideoModelV3 as Experimental_VideoModelV3, type VideoModelV3CallOptions as Experimental_VideoModelV3CallOptions, type VideoModelV3File as Experimental_VideoModelV3File, type VideoModelV3VideoData as Experimental_VideoModelV3VideoData, type ImageModelV2, type ImageModelV2CallOptions, type ImageModelV2CallWarning, type ImageModelV2ProviderMetadata, type ImageModelV3, type ImageModelV3CallOptions, type ImageModelV3File, type ImageModelV3Middleware, type ImageModelV3ProviderMetadata, type ImageModelV3Usage, InvalidArgumentError, InvalidPromptError, InvalidResponseDataError, type JSONArray, type JSONObject, JSONParseError, type JSONValue, type LanguageModelV2, type LanguageModelV2CallOptions, type LanguageModelV2CallWarning, type LanguageModelV2Content, type LanguageModelV2DataContent, type LanguageModelV2File, type LanguageModelV2FilePart, type LanguageModelV2FinishReason, type LanguageModelV2FunctionTool, type LanguageModelV2Message, type LanguageModelV2Middleware, type LanguageModelV2Prompt, type LanguageModelV2ProviderDefinedTool, type LanguageModelV2Reasoning, type LanguageModelV2ReasoningPart, type LanguageModelV2ResponseMetadata, type LanguageModelV2Source, type LanguageModelV2StreamPart, type LanguageModelV2Text, type LanguageModelV2TextPart, type LanguageModelV2ToolCall, type LanguageModelV2ToolCallPart, type LanguageModelV2ToolChoice, type LanguageModelV2ToolResultOutput, type LanguageModelV2ToolResultPart, type LanguageModelV2Usage, type LanguageModelV3, type LanguageModelV3CallOptions, type LanguageModelV3Content, type LanguageModelV3DataContent, type LanguageModelV3File, type LanguageModelV3FilePart, type LanguageModelV3FinishReason, type LanguageModelV3FunctionTool, type LanguageModelV3GenerateResult, type LanguageModelV3Message, type LanguageModelV3Middleware, type LanguageModelV3Prompt, type LanguageModelV3ProviderTool, type LanguageModelV3Reasoning, type LanguageModelV3ReasoningPart, type LanguageModelV3ResponseMetadata, type LanguageModelV3Source, type LanguageModelV3StreamPart, type LanguageModelV3StreamResult, type LanguageModelV3Text, type LanguageModelV3TextPart, type LanguageModelV3ToolApprovalRequest, type LanguageModelV3ToolApprovalResponsePart, type LanguageModelV3ToolCall, type LanguageModelV3ToolCallPart, type LanguageModelV3ToolChoice, type LanguageModelV3ToolResult, type LanguageModelV3ToolResultOutput, type LanguageModelV3ToolResultPart, type LanguageModelV3Usage, LoadAPIKeyError, LoadSettingError, NoContentGeneratedError, NoSuchModelError, type ProviderV2, type ProviderV3, type RerankingModelV3, type RerankingModelV3CallOptions, type SharedV2Headers, type SharedV2ProviderMetadata, type SharedV2ProviderOptions, type SharedV3Headers, type SharedV3ProviderMetadata, type SharedV3ProviderOptions, type SharedV3Warning, type SpeechModelV2, type SpeechModelV2CallOptions, type SpeechModelV2CallWarning, type SpeechModelV3, type SpeechModelV3CallOptions, TooManyEmbeddingValuesForCallError, type TranscriptionModelV2, type TranscriptionModelV2CallOptions, type TranscriptionModelV2CallWarning, type TranscriptionModelV3, type TranscriptionModelV3CallOptions, type TypeValidationContext, TypeValidationError, UnsupportedFunctionalityError, getErrorMessage, isJSONArray, isJSONObject, isJSONValue };
+export { AISDKError, APICallError, type EmbeddingModelV2, type EmbeddingModelV2Embedding, type EmbeddingModelV3, type EmbeddingModelV3CallOptions, type EmbeddingModelV3Embedding, type EmbeddingModelV3Middleware, type EmbeddingModelV3Result, EmptyResponseBodyError, type VideoModelV3 as Experimental_VideoModelV3, type VideoModelV3CallOptions as Experimental_VideoModelV3CallOptions, type VideoModelV3File as Experimental_VideoModelV3File, type VideoModelV3FrameImage as Experimental_VideoModelV3FrameImage, type VideoModelV3FrameType as Experimental_VideoModelV3FrameType, type VideoModelV3VideoData as Experimental_VideoModelV3VideoData, type ImageModelV2, type ImageModelV2CallOptions, type ImageModelV2CallWarning, type ImageModelV2ProviderMetadata, type ImageModelV3, type ImageModelV3CallOptions, type ImageModelV3File, type ImageModelV3Middleware, type ImageModelV3ProviderMetadata, type ImageModelV3Usage, InvalidArgumentError, InvalidPromptError, InvalidResponseDataError, type JSONArray, type JSONObject, JSONParseError, type JSONValue, type LanguageModelV2, type LanguageModelV2CallOptions, type LanguageModelV2CallWarning, type LanguageModelV2Content, type LanguageModelV2DataContent, type LanguageModelV2File, type LanguageModelV2FilePart, type LanguageModelV2FinishReason, type LanguageModelV2FunctionTool, type LanguageModelV2Message, type LanguageModelV2Middleware, type LanguageModelV2Prompt, type LanguageModelV2ProviderDefinedTool, type LanguageModelV2Reasoning, type LanguageModelV2ReasoningPart, type LanguageModelV2ResponseMetadata, type LanguageModelV2Source, type LanguageModelV2StreamPart, type LanguageModelV2Text, type LanguageModelV2TextPart, type LanguageModelV2ToolCall, type LanguageModelV2ToolCallPart, type LanguageModelV2ToolChoice, type LanguageModelV2ToolResultOutput, type LanguageModelV2ToolResultPart, type LanguageModelV2Usage, type LanguageModelV3, type LanguageModelV3CallOptions, type LanguageModelV3Content, type LanguageModelV3DataContent, type LanguageModelV3File, type LanguageModelV3FilePart, type LanguageModelV3FinishReason, type LanguageModelV3FunctionTool, type LanguageModelV3GenerateResult, type LanguageModelV3Message, type LanguageModelV3Middleware, type LanguageModelV3Prompt, type LanguageModelV3ProviderTool, type LanguageModelV3Reasoning, type LanguageModelV3ReasoningPart, type LanguageModelV3ResponseMetadata, type LanguageModelV3Source, type LanguageModelV3StreamPart, type LanguageModelV3StreamResult, type LanguageModelV3Text, type LanguageModelV3TextPart, type LanguageModelV3ToolApprovalRequest, type LanguageModelV3ToolApprovalResponsePart, type LanguageModelV3ToolCall, type LanguageModelV3ToolCallPart, type LanguageModelV3ToolChoice, type LanguageModelV3ToolResult, type LanguageModelV3ToolResultOutput, type LanguageModelV3ToolResultPart, type LanguageModelV3Usage, LoadAPIKeyError, LoadSettingError, NoContentGeneratedError, NoSuchModelError, type ProviderV2, type ProviderV3, type RerankingModelV3, type RerankingModelV3CallOptions, type SharedV2Headers, type SharedV2ProviderMetadata, type SharedV2ProviderOptions, type SharedV3Headers, type SharedV3ProviderMetadata, type SharedV3ProviderOptions, type SharedV3Warning, type SpeechModelV2, type SpeechModelV2CallOptions, type SpeechModelV2CallWarning, type SpeechModelV3, type SpeechModelV3CallOptions, TooManyEmbeddingValuesForCallError, type TranscriptionModelV2, type TranscriptionModelV2CallOptions, type TranscriptionModelV2CallWarning, type TranscriptionModelV3, type TranscriptionModelV3CallOptions, type TypeValidationContext, TypeValidationError, UnsupportedFunctionalityError, getErrorMessage, isJSONArray, isJSONObject, isJSONValue };

package/dist/index.d.ts

@@ -3843,6 +3843,27 @@ type VideoModelV3File = {
     providerOptions?: SharedV3ProviderMetadata;
 };
 
+/**
+ * The role a frame image plays in video generation.
+ *
+ * - `first_frame`: the starting frame the model animates from
+ * - `last_frame`: the ending frame the model animates towards
+ */
+type VideoModelV3FrameType = 'first_frame' | 'last_frame';
+/**
+ * A role-tagged image input for image-to-video and first-last-frame generation.
+ */
+type VideoModelV3FrameImage = {
+    /**
+     * The image file used for this frame.
+     */
+    image: VideoModelV3File;
+    /**
+     * Which frame this image represents.
+     */
+    frameType: VideoModelV3FrameType;
+};
+
 type VideoModelV3CallOptions = {
     /**
      * Text prompt for the video generation.
@@ -3888,6 +3909,16 @@ type VideoModelV3CallOptions = {
      * The image serves as the starting frame that the model will animate.
      */
     image: VideoModelV3File | undefined;
+    /**
+     * Role-tagged image inputs for first-last-frame generation.
+     * Each entry declares whether it is the `first_frame` or the
+     * `last_frame` of the generated video.
+     */
+    frameImages: Array<VideoModelV3FrameImage> | undefined;
+    /**
+     * Reference image inputs for reference-to-video generation.
+     */
+    inputReferences: Array<VideoModelV3File> | undefined;
     /**
      * Whether the model should generate audio alongside the video.
      */
@@ -4031,4 +4062,4 @@ type VideoModelV3 = {
     }>;
 };
 
-export { AISDKError, APICallError, type EmbeddingModelV2, type EmbeddingModelV2Embedding, type EmbeddingModelV3, type EmbeddingModelV3CallOptions, type EmbeddingModelV3Embedding, type EmbeddingModelV3Middleware, type EmbeddingModelV3Result, EmptyResponseBodyError, type VideoModelV3 as Experimental_VideoModelV3, type VideoModelV3CallOptions as Experimental_VideoModelV3CallOptions, type VideoModelV3File as Experimental_VideoModelV3File, type VideoModelV3VideoData as Experimental_VideoModelV3VideoData, type ImageModelV2, type ImageModelV2CallOptions, type ImageModelV2CallWarning, type ImageModelV2ProviderMetadata, type ImageModelV3, type ImageModelV3CallOptions, type ImageModelV3File, type ImageModelV3Middleware, type ImageModelV3ProviderMetadata, type ImageModelV3Usage, InvalidArgumentError, InvalidPromptError, InvalidResponseDataError, type JSONArray, type JSONObject, JSONParseError, type JSONValue, type LanguageModelV2, type LanguageModelV2CallOptions, type LanguageModelV2CallWarning, type LanguageModelV2Content, type LanguageModelV2DataContent, type LanguageModelV2File, type LanguageModelV2FilePart, type LanguageModelV2FinishReason, type LanguageModelV2FunctionTool, type LanguageModelV2Message, type LanguageModelV2Middleware, type LanguageModelV2Prompt, type LanguageModelV2ProviderDefinedTool, type LanguageModelV2Reasoning, type LanguageModelV2ReasoningPart, type LanguageModelV2ResponseMetadata, type LanguageModelV2Source, type LanguageModelV2StreamPart, type LanguageModelV2Text, type LanguageModelV2TextPart, type LanguageModelV2ToolCall, type LanguageModelV2ToolCallPart, type LanguageModelV2ToolChoice, type LanguageModelV2ToolResultOutput, type LanguageModelV2ToolResultPart, type LanguageModelV2Usage, type LanguageModelV3, type LanguageModelV3CallOptions, type LanguageModelV3Content, type LanguageModelV3DataContent, type LanguageModelV3File, type LanguageModelV3FilePart, type LanguageModelV3FinishReason, type LanguageModelV3FunctionTool, type LanguageModelV3GenerateResult, type LanguageModelV3Message, type LanguageModelV3Middleware, type LanguageModelV3Prompt, type LanguageModelV3ProviderTool, type LanguageModelV3Reasoning, type LanguageModelV3ReasoningPart, type LanguageModelV3ResponseMetadata, type LanguageModelV3Source, type LanguageModelV3StreamPart, type LanguageModelV3StreamResult, type LanguageModelV3Text, type LanguageModelV3TextPart, type LanguageModelV3ToolApprovalRequest, type LanguageModelV3ToolApprovalResponsePart, type LanguageModelV3ToolCall, type LanguageModelV3ToolCallPart, type LanguageModelV3ToolChoice, type LanguageModelV3ToolResult, type LanguageModelV3ToolResultOutput, type LanguageModelV3ToolResultPart, type LanguageModelV3Usage, LoadAPIKeyError, LoadSettingError, NoContentGeneratedError, NoSuchModelError, type ProviderV2, type ProviderV3, type RerankingModelV3, type RerankingModelV3CallOptions, type SharedV2Headers, type SharedV2ProviderMetadata, type SharedV2ProviderOptions, type SharedV3Headers, type SharedV3ProviderMetadata, type SharedV3ProviderOptions, type SharedV3Warning, type SpeechModelV2, type SpeechModelV2CallOptions, type SpeechModelV2CallWarning, type SpeechModelV3, type SpeechModelV3CallOptions, TooManyEmbeddingValuesForCallError, type TranscriptionModelV2, type TranscriptionModelV2CallOptions, type TranscriptionModelV2CallWarning, type TranscriptionModelV3, type TranscriptionModelV3CallOptions, type TypeValidationContext, TypeValidationError, UnsupportedFunctionalityError, getErrorMessage, isJSONArray, isJSONObject, isJSONValue };
+export { AISDKError, APICallError, type EmbeddingModelV2, type EmbeddingModelV2Embedding, type EmbeddingModelV3, type EmbeddingModelV3CallOptions, type EmbeddingModelV3Embedding, type EmbeddingModelV3Middleware, type EmbeddingModelV3Result, EmptyResponseBodyError, type VideoModelV3 as Experimental_VideoModelV3, type VideoModelV3CallOptions as Experimental_VideoModelV3CallOptions, type VideoModelV3File as Experimental_VideoModelV3File, type VideoModelV3FrameImage as Experimental_VideoModelV3FrameImage, type VideoModelV3FrameType as Experimental_VideoModelV3FrameType, type VideoModelV3VideoData as Experimental_VideoModelV3VideoData, type ImageModelV2, type ImageModelV2CallOptions, type ImageModelV2CallWarning, type ImageModelV2ProviderMetadata, type ImageModelV3, type ImageModelV3CallOptions, type ImageModelV3File, type ImageModelV3Middleware, type ImageModelV3ProviderMetadata, type ImageModelV3Usage, InvalidArgumentError, InvalidPromptError, InvalidResponseDataError, type JSONArray, type JSONObject, JSONParseError, type JSONValue, type LanguageModelV2, type LanguageModelV2CallOptions, type LanguageModelV2CallWarning, type LanguageModelV2Content, type LanguageModelV2DataContent, type LanguageModelV2File, type LanguageModelV2FilePart, type LanguageModelV2FinishReason, type LanguageModelV2FunctionTool, type LanguageModelV2Message, type LanguageModelV2Middleware, type LanguageModelV2Prompt, type LanguageModelV2ProviderDefinedTool, type LanguageModelV2Reasoning, type LanguageModelV2ReasoningPart, type LanguageModelV2ResponseMetadata, type LanguageModelV2Source, type LanguageModelV2StreamPart, type LanguageModelV2Text, type LanguageModelV2TextPart, type LanguageModelV2ToolCall, type LanguageModelV2ToolCallPart, type LanguageModelV2ToolChoice, type LanguageModelV2ToolResultOutput, type LanguageModelV2ToolResultPart, type LanguageModelV2Usage, type LanguageModelV3, type LanguageModelV3CallOptions, type LanguageModelV3Content, type LanguageModelV3DataContent, type LanguageModelV3File, type LanguageModelV3FilePart, type LanguageModelV3FinishReason, type LanguageModelV3FunctionTool, type LanguageModelV3GenerateResult, type LanguageModelV3Message, type LanguageModelV3Middleware, type LanguageModelV3Prompt, type LanguageModelV3ProviderTool, type LanguageModelV3Reasoning, type LanguageModelV3ReasoningPart, type LanguageModelV3ResponseMetadata, type LanguageModelV3Source, type LanguageModelV3StreamPart, type LanguageModelV3StreamResult, type LanguageModelV3Text, type LanguageModelV3TextPart, type LanguageModelV3ToolApprovalRequest, type LanguageModelV3ToolApprovalResponsePart, type LanguageModelV3ToolCall, type LanguageModelV3ToolCallPart, type LanguageModelV3ToolChoice, type LanguageModelV3ToolResult, type LanguageModelV3ToolResultOutput, type LanguageModelV3ToolResultPart, type LanguageModelV3Usage, LoadAPIKeyError, LoadSettingError, NoContentGeneratedError, NoSuchModelError, type ProviderV2, type ProviderV3, type RerankingModelV3, type RerankingModelV3CallOptions, type SharedV2Headers, type SharedV2ProviderMetadata, type SharedV2ProviderOptions, type SharedV3Headers, type SharedV3ProviderMetadata, type SharedV3ProviderOptions, type SharedV3Warning, type SpeechModelV2, type SpeechModelV2CallOptions, type SpeechModelV2CallWarning, type SpeechModelV3, type SpeechModelV3CallOptions, TooManyEmbeddingValuesForCallError, type TranscriptionModelV2, type TranscriptionModelV2CallOptions, type TranscriptionModelV2CallWarning, type TranscriptionModelV3, type TranscriptionModelV3CallOptions, type TypeValidationContext, TypeValidationError, UnsupportedFunctionalityError, getErrorMessage, isJSONArray, isJSONObject, isJSONValue };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ai-sdk/provider",
-  "version": "3.0.12",
+  "version": "3.0.13",
   "license": "Apache-2.0",
   "sideEffects": false,
   "main": "./dist/index.js",

package/src/video-model/v3/index.ts

@@ -4,3 +4,7 @@ export type {
 } from './video-model-v3';
 export type { VideoModelV3CallOptions as Experimental_VideoModelV3CallOptions } from './video-model-v3-call-options';
 export type { VideoModelV3File as Experimental_VideoModelV3File } from './video-model-v3-file';
+export type {
+  VideoModelV3FrameImage as Experimental_VideoModelV3FrameImage,
+  VideoModelV3FrameType as Experimental_VideoModelV3FrameType,
+} from './video-model-v3-frame-image';

package/src/video-model/v3/video-model-v3-call-options.ts

@@ -1,5 +1,6 @@
 import type { SharedV3ProviderOptions } from '../../shared';
 import type { VideoModelV3File } from './video-model-v3-file';
+import type { VideoModelV3FrameImage } from './video-model-v3-frame-image';
 
 export type VideoModelV3CallOptions = {
   /**
@@ -54,6 +55,18 @@ export type VideoModelV3CallOptions = {
    */
   image: VideoModelV3File | undefined;
 
+  /**
+   * Role-tagged image inputs for first-last-frame generation.
+   * Each entry declares whether it is the `first_frame` or the
+   * `last_frame` of the generated video.
+   */
+  frameImages: Array<VideoModelV3FrameImage> | undefined;
+
+  /**
+   * Reference image inputs for reference-to-video generation.
+   */
+  inputReferences: Array<VideoModelV3File> | undefined;
+
   /**
    * Whether the model should generate audio alongside the video.
    */

package/src/video-model/v3/video-model-v3-frame-image.ts

@@ -0,0 +1,24 @@
+import type { VideoModelV3File } from './video-model-v3-file';
+
+/**
+ * The role a frame image plays in video generation.
+ *
+ * - `first_frame`: the starting frame the model animates from
+ * - `last_frame`: the ending frame the model animates towards
+ */
+export type VideoModelV3FrameType = 'first_frame' | 'last_frame';
+
+/**
+ * A role-tagged image input for image-to-video and first-last-frame generation.
+ */
+export type VideoModelV3FrameImage = {
+  /**
+   * The image file used for this frame.
+   */
+  image: VideoModelV3File;
+
+  /**
+   * Which frame this image represents.
+   */
+  frameType: VideoModelV3FrameType;
+};