diff --git a/src/app/api/generate/providers/kie.ts b/src/app/api/generate/providers/kie.ts
index f85e13c0..87fe2191 100644
--- a/src/app/api/generate/providers/kie.ts
+++ b/src/app/api/generate/providers/kie.ts
@@ -624,9 +624,12 @@ export async function submitKieTask(
     }
   }
 
-  // Handle image inputs (fallback - only if dynamicInputs didn't already set the image key)
+  // Handle image inputs (fallback - only if dynamicInputs didn't already handle any image key).
+  // If the schema-driven dynamic inputs already placed images into specific fields
+  // (e.g. reference_image_urls), don't also populate the default key — some models
+  // treat these fields as mutually exclusive and reject the request.
   const imageKey = getKieImageInputKey(modelId);
-  if (input.images && input.images.length > 0 && !handledImageKeys.has(imageKey)) {
+  if (input.images && input.images.length > 0 && handledImageKeys.size === 0) {
     const imageUrls: string[] = [];
     for (const image of input.images) {
       if (image.startsWith("http")) {
diff --git a/src/app/api/models/[modelId]/route.ts b/src/app/api/models/[modelId]/route.ts
index d8354bab..d4d7d83c 100644
--- a/src/app/api/models/[modelId]/route.ts
+++ b/src/app/api/models/[modelId]/route.ts
@@ -881,10 +881,10 @@ function getKieSchema(modelId: string): ExtractedSchema {
       ],
       inputs: [
         { name: "prompt", type: "text", required: false, label: "Prompt" },
-        { name: "first_frame_url", type: "image", required: true, label: "First Frame", description: "Starting frame of the generated video." },
-        { name: "last_frame_url", type: "image", required: false, label: "Last Frame", description: "Optional end frame — the video interpolates between first and last." },
-        { name: "reference_image_urls", type: "image", required: false, isArray: true, label: "Reference Images", description: "Up to 9 supplementary reference images for style/subject guidance." },
-        { name: "reference_video_urls", type: "image", required: false, isArray: true, label: "Reference Videos", description: "Up to 3 reference videos (uses image handle; accepts video data URLs)." },
+        { name: "first_frame_url", type: "image", required: false, label: "First Frame", description: "Starting frame. Mutually exclusive with Reference Images." },
+        { name: "last_frame_url", type: "image", required: false, label: "Last Frame", description: "Optional end frame — interpolates from First to Last. Mutually exclusive with Reference Images." },
+        { name: "reference_image_urls", type: "image", required: false, isArray: true, label: "Reference Images", description: "Up to 9 reference images for style/subject guidance. Mutually exclusive with First/Last Frame." },
+        { name: "reference_video_urls", type: "image", required: false, isArray: true, label: "Reference Videos", description: "Up to 3 reference videos (accepts video data URLs over the image handle)." },
         { name: "reference_audio_urls", type: "audio", required: false, isArray: true, label: "Reference Audio", description: "Up to 3 reference audio clips." },
       ],
     },
@@ -910,10 +910,10 @@ function getKieSchema(modelId: string): ExtractedSchema {
       ],
       inputs: [
         { name: "prompt", type: "text", required: false, label: "Prompt" },
-        { name: "first_frame_url", type: "image", required: true, label: "First Frame", description: "Starting frame of the generated video." },
-        { name: "last_frame_url", type: "image", required: false, label: "Last Frame", description: "Optional end frame — the video interpolates between first and last." },
-        { name: "reference_image_urls", type: "image", required: false, isArray: true, label: "Reference Images", description: "Up to 9 supplementary reference images for style/subject guidance." },
-        { name: "reference_video_urls", type: "image", required: false, isArray: true, label: "Reference Videos", description: "Up to 3 reference videos (uses image handle; accepts video data URLs)." },
+        { name: "first_frame_url", type: "image", required: false, label: "First Frame", description: "Starting frame. Mutually exclusive with Reference Images." },
+        { name: "last_frame_url", type: "image", required: false, label: "Last Frame", description: "Optional end frame — interpolates from First to Last. Mutually exclusive with Reference Images." },
+        { name: "reference_image_urls", type: "image", required: false, isArray: true, label: "Reference Images", description: "Up to 9 reference images for style/subject guidance. Mutually exclusive with First/Last Frame." },
+        { name: "reference_video_urls", type: "image", required: false, isArray: true, label: "Reference Videos", description: "Up to 3 reference videos (accepts video data URLs over the image handle)." },
         { name: "reference_audio_urls", type: "audio", required: false, isArray: true, label: "Reference Audio", description: "Up to 3 reference audio clips." },
       ],
     },