feat: add audio-to-video capability type and model discovery

Add "audio-to-video" to ModelCapability union, extend ModelInput.type
and ModelInputDef.type with "audio", and include "audio-to-video" in
fal.ai RELEVANT_CATEGORIES for model discovery.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

src/app/api/models/route.ts

@@ -54,6 +54,7 @@ const RELEVANT_CATEGORIES = [
   "text-to-speech",
   "text-to-music",
   "text-to-sound-effects",
+  "audio-to-video",
 ];
 
 // Kie.ai models (hardcoded - no discovery API available)

src/lib/providers/types.ts

@@ -18,7 +18,8 @@ export type ModelCapability =
   | "image-to-video"
   | "text-to-3d"
   | "image-to-3d"
-  | "text-to-audio";
+  | "text-to-audio"
+  | "audio-to-video";
 
 /**
  * Model parameter schema for dynamic UI generation
@@ -42,7 +43,7 @@ export interface ModelInput {
   /** Property name from schema (e.g., "image_url", "tail_image_url", "prompt") */
   name: string;
   /** Handle type for connections */
-  type: "image" | "text";
+  type: "image" | "text" | "audio";
   /** Whether this input is required */
   required: boolean;
   /** Human-readable label for the handle */

src/types/nodes.ts

@@ -172,7 +172,7 @@ export interface CarouselVideoItem {
  */
 export interface ModelInputDef {
   name: string;
-  type: "image" | "text";
+  type: "image" | "text" | "audio";
   required: boolean;
   label: string;
   description?: string;