fal-endpoint-types 1.3.8 → 1.3.10

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "fal-endpoint-types",
-    "version": "1.3.8",
+    "version": "1.3.10",
     "description": "TypeScript types for Fal AI endpoints generated from the OpenAPI schemas.",
     "homepage": "https://github.com/rawpixel-vincent/fal-endpoint-types#readme",
     "bugs": {

package/types/fal/endpoints/components.d.ts

@@ -630,6 +630,60 @@ export interface TranscriptionWord {
     type: string;
 }
 
+export interface TrajectoryParameters {
+    /**
+     * Phi
+     * @description Vertical rotation angles (degrees) for each keyframe.
+     * @example [
+     *       0,
+     *       -3,
+     *       -8,
+     *       -15,
+     *       -20,
+     *       -15,
+     *       -10,
+     *       -5,
+     *       0
+     *     ]
+     */
+    phi: number[];
+    /**
+     * Radius
+     * @description Camera distance scaling factors for each keyframe.
+     * @example [
+     *       0,
+     *       0.02,
+     *       0.09,
+     *       0.16,
+     *       0.25,
+     *       0.2,
+     *       0.09,
+     *       0
+     *     ]
+     */
+    radius: number[];
+    /**
+     * Theta
+     * @description Horizontal rotation angles (degrees) for each keyframe.
+     * @example [
+     *       0,
+     *       2,
+     *       8,
+     *       10,
+     *       5,
+     *       3,
+     *       0,
+     *       -2,
+     *       -5,
+     *       -8,
+     *       -5,
+     *       -3,
+     *       0
+     *     ]
+     */
+    theta: number[];
+}
+
 export interface Trajectory {
     /**
      * X
@@ -801,6 +855,19 @@ export interface StructuredInstruction {
     text_render?: { [x: string]: any }[];
 }
 
+export interface SpeechTimestamp {
+    /**
+     * End Time
+     * @description The end time of the speech in seconds.
+     */
+    end: number;
+    /**
+     * Start Time
+     * @description The start time of the speech in seconds.
+     */
+    start: number;
+}
+
 export interface Speaker {
     /** Audio Url */
     audio_url: string;
@@ -841,80 +908,6 @@ export interface SemanticImageInput {
     reference: string;
 }
 
-export interface SAM3DObjectMetadata {
-    /**
-     * Camera Pose
-     * @description Camera pose matrix
-     */
-    camera_pose?: number[][];
-    /**
-     * Object Index
-     * @description Index of the object in the scene
-     */
-    object_index: number;
-    /**
-     * Rotation
-     * @description Rotation quaternion [x, y, z, w]
-     */
-    rotation?: number[][];
-    /**
-     * Scale
-     * @description Scale factors [sx, sy, sz]
-     */
-    scale?: number[][];
-    /**
-     * Translation
-     * @description Translation [tx, ty, tz]
-     */
-    translation?: number[][];
-}
-
-export interface SAM3DBodyPersonMetadata {
-    /**
-     * Bbox
-     * @description Bounding box [x_min, y_min, x_max, y_max]
-     */
-    bbox: number[];
-    /**
-     * Focal Length
-     * @description Estimated focal length
-     */
-    focal_length: number;
-    /**
-     * Keypoints 2D
-     * @description 2D keypoints [[x, y], ...] - 70 body keypoints
-     */
-    keypoints_2d: number[][];
-    /**
-     * Keypoints 3D
-     * @description 3D keypoints [[x, y, z], ...] - 70 body keypoints in camera space
-     */
-    keypoints_3d?: number[][];
-    /**
-     * Person Id
-     * @description Index of the person in the scene
-     */
-    person_id: number;
-    /**
-     * Pred Cam T
-     * @description Predicted camera translation [tx, ty, tz]
-     */
-    pred_cam_t: number[];
-}
-
-export interface SAM3DBodyMetadata {
-    /**
-     * Num People
-     * @description Number of people detected
-     */
-    num_people: number;
-    /**
-     * People
-     * @description Per-person metadata
-     */
-    people: SAM3DBodyPersonMetadata[];
-}
-
 export interface SAM3DBodyAlignmentInfo {
     /**
      * Cropped Vertices Count
@@ -1310,7 +1303,7 @@ export interface OmniVideoElementInput {
     frontal_image_url: string;
     /**
      * Reference Image Urls
-     * @description Additional reference images from different angles. 1-3 images supported. At least one image is required.
+     * @description Additional reference images from different angles. 1-4 images supported. At least one image is required.
      */
     reference_image_urls?: string[];
 }
@@ -1495,6 +1488,24 @@ export interface ModelUrls {
     usdz?: File_1;
 }
 
+export interface MaskMetadata {
+    /**
+     * Box
+     * @description Bounding box for the mask in normalized cxcywh coordinates.
+     */
+    box?: number[];
+    /**
+     * Index
+     * @description Index of the mask inside the model output.
+     */
+    index: number;
+    /**
+     * Score
+     * @description Score for this mask.
+     */
+    score?: number;
+}
+
 export interface LoudnormSummary {
     /**
      * Input Integrated
@@ -2410,58 +2421,6 @@ export interface File {
     url: string;
 }
 
-export interface ExtendVideoConditioningInput {
-    /**
-     * Limit Number of Frames
-     * @description Whether to limit the number of frames used from the video. If True, the `max_num_frames` parameter will be used to limit the number of frames.
-     * @default false
-     */
-    limit_num_frames?: boolean;
-    /**
-     * Maximum Number of Frames
-     * @description Maximum number of frames to use from the video. If None, all frames will be used.
-     * @default 1441
-     * @example 1441
-     */
-    max_num_frames?: number;
-    /**
-     * Resample FPS
-     * @description Whether to resample the video to a specific FPS. If True, the `target_fps` parameter will be used to resample the video.
-     * @default false
-     */
-    resample_fps?: boolean;
-    /**
-     * Reverse Video
-     * @description Whether to reverse the video. This is useful for tasks where the video conditioning should be applied in reverse order.
-     * @default false
-     */
-    reverse_video?: boolean;
-    /**
-     * Start Frame Number
-     * @description Frame number of the video from which the conditioning starts. Must be a multiple of 8.
-     * @default 0
-     */
-    start_frame_num?: number;
-    /**
-     * Strength
-     * @description Strength of the conditioning. 0.0 means no conditioning, 1.0 means full conditioning.
-     * @default 1
-     */
-    strength?: number;
-    /**
-     * Target FPS
-     * @description Target FPS to resample the video to. Only relevant if `resample_fps` is True.
-     * @default 24
-     * @example 24
-     */
-    target_fps?: number;
-    /**
-     * Video URL
-     * @description URL of video to use as conditioning
-     */
-    video_url: string;
-}
-
 export interface EmotionalStrengths {
     /**
      * Afraid
@@ -2744,38 +2703,6 @@ export interface ControlNetUnion {
     variant?: string;
 }
 
-export interface ControlNet_2 {
-    /**
-     * Conditioning Scale
-     * @description The scale of the control net weight. This is used to scale the control net weight
-     *                 before merging it with the base model.
-     * @default 1
-     */
-    conditioning_scale?: number;
-    /**
-     * Control Image Url
-     * @description URL of the image to be used as the control image.
-     */
-    control_image_url: string;
-    /**
-     * End Percentage
-     * @description The percentage of the image to end applying the controlnet in terms of the total timesteps.
-     * @default 1
-     */
-    end_percentage?: number;
-    /**
-     * Path
-     * @description URL or the path to the control net weights.
-     */
-    path: string;
-    /**
-     * Start Percentage
-     * @description The percentage of the image to start applying the controlnet in terms of the total timesteps.
-     * @default 0
-     */
-    start_percentage?: number;
-}
-
 export interface ControlNet_1 {
     /**
      * Conditioning Scale
@@ -2908,6 +2835,24 @@ export interface ControlLoraWeight {
     scale?: Record<string, number> | number;
 }
 
+export interface CompletionUsage {
+    /**
+     * Completion Tokens
+     * @description Number of tokens in the completion
+     */
+    completion_tokens: number;
+    /**
+     * Prompt Tokens
+     * @description Number of tokens in the prompt
+     */
+    prompt_tokens: number;
+    /**
+     * Total Tokens
+     * @description Total tokens used
+     */
+    total_tokens: number;
+}
+
 export interface ColorPaletteMember {
     /**
      * Color Weight
@@ -3037,56 +2982,56 @@ export interface bria_fibovlm_Aesthetics {
 }
 
 export interface BoxPromptBase_1 {
+    /**
+     * Object Id
+     * @description Optional object identifier. Boxes sharing an object id refine the same object.
+     */
+    object_id?: number;
     /**
      * X Max
-     * @description X Max Coordinate of the prompt
-     * @default 0
+     * @description X Max Coordinate of the box
      */
     x_max?: number;
     /**
      * X Min
      * @description X Min Coordinate of the box
-     * @default 0
      */
     x_min?: number;
     /**
      * Y Max
-     * @description Y Max Coordinate of the prompt
-     * @default 0
+     * @description Y Max Coordinate of the box
      */
     y_max?: number;
     /**
      * Y Min
      * @description Y Min Coordinate of the box
-     * @default 0
      */
     y_min?: number;
 }
 
 export interface BoxPromptBase {
-    /**
-     * Object Id
-     * @description Optional object identifier. Boxes sharing an object id refine the same object.
-     */
-    object_id?: number;
     /**
      * X Max
-     * @description X Max Coordinate of the box
+     * @description X Max Coordinate of the prompt
+     * @default 0
      */
     x_max?: number;
     /**
      * X Min
      * @description X Min Coordinate of the box
+     * @default 0
      */
     x_min?: number;
     /**
      * Y Max
-     * @description Y Max Coordinate of the box
+     * @description Y Max Coordinate of the prompt
+     * @default 0
      */
     y_max?: number;
     /**
      * Y Min
      * @description Y Min Coordinate of the box
+     * @default 0
      */
     y_min?: number;
 }
@@ -3321,7 +3266,7 @@ export interface AudioSetting {
     sample_rate?: 8000 | 16000 | 22050 | 24000 | 32000 | 44100;
 }
 
-export interface AudioFile_1 {
+export interface AudioFile_2 {
     /**
      * Content Type
      * @default audio/wav
@@ -3347,7 +3292,7 @@ export interface AudioFile_1 {
     url: string;
 }
 
-export interface AudioFile {
+export interface AudioFile_1 {
     /**
      * Bitrate
      * @description The bitrate of the audio
@@ -3399,6 +3344,58 @@ export interface AudioFile {
     url: string;
 }
 
+export interface AudioFile {
+    /**
+     * Bitrate
+     * @description The bitrate of the audio (e.g., '192k' or 192000)
+     */
+    bitrate?: string | number;
+    /**
+     * Channels
+     * @description The number of channels in the audio
+     */
+    channels?: number;
+    /**
+     * Content Type
+     * @description The mime type of the file.
+     * @example image/png
+     */
+    content_type?: string;
+    /**
+     * Duration
+     * @description The duration of the audio
+     */
+    duration?: number;
+    /**
+     * File Data
+     * Format: binary
+     * @description File data
+     */
+    file_data?: string;
+    /**
+     * File Name
+     * @description The name of the file. It will be auto-generated if not provided.
+     * @example z9RV14K95DvU.png
+     */
+    file_name?: string;
+    /**
+     * File Size
+     * @description The size of the file in bytes.
+     * @example 4404019
+     */
+    file_size?: number;
+    /**
+     * Sample Rate
+     * @description The sample rate of the audio
+     */
+    sample_rate?: number;
+    /**
+     * Url
+     * @description The URL where the file can be downloaded from.
+     */
+    url: string;
+}
+
 export interface Audio {
     /**
      * Bitrate