From d4b49df05a36b12d40a34c0837603d1f172f0db7 Mon Sep 17 00:00:00 2001
From: Towsty <ianjmcarr@gmail.com>
Date: Sun, 6 Sep 2026 07:35:42 -0500
Subject: [PATCH] Let text-to-video generate without a still: Still/Text lives
 on the Video card, and Text mode skips the required input image.

Co-authored-by: Cursor <cursoragent@cursor.com>
---
 components/StudioKindCards.vue | 28 ++++++++++++++
 pages/index.vue                | 71 ++++++++++++++++------------------
 2 files changed, 62 insertions(+), 37 deletions(-)

diff --git a/components/StudioKindCards.vue b/components/StudioKindCards.vue
index dabe601..ff2a823 100644
--- a/components/StudioKindCards.vue
+++ b/components/StudioKindCards.vue
@@ -38,6 +38,26 @@
         </button>
       </div>
       <p v-else class="border-t border-white/10 px-4 py-2 text-xs text-zinc-500">MiniMax H3</p>
+      <div class="grid grid-cols-2 border-t border-white/10">
+        <button
+          type="button"
+          class="border-r border-white/10 px-3 py-2.5 text-left text-sm transition"
+          :class="engineBtnClass(current === 'video' && videoStart === 'still')"
+          @click="pickStart('still')"
+        >
+          <span class="block font-semibold">Still</span>
+          <span class="text-xs opacity-80">Image to video</span>
+        </button>
+        <button
+          type="button"
+          class="px-3 py-2.5 text-left text-sm transition"
+          :class="engineBtnClass(current === 'video' && videoStart === 'text')"
+          @click="pickStart('text')"
+        >
+          <span class="block font-semibold">Text</span>
+          <span class="text-xs opacity-80">No start still</span>
+        </button>
+      </div>
     </div>
 
     <div
@@ -123,11 +143,13 @@ const props = withDefaults(defineProps<{
   current: 'video' | 'image' | 'music'
   ltxEnabled?: boolean
   videoEngine?: 'minimax' | 'ltx'
+  videoStart?: 'still' | 'text'
   imageEngine?: 'flux' | 'krea'
   musicEngine?: 'ace' | 'yue'
 }>(), {
   ltxEnabled: false,
   videoEngine: 'minimax',
+  videoStart: 'still',
   imageEngine: 'flux',
   musicEngine: 'ace'
 })
@@ -135,6 +157,7 @@ const props = withDefaults(defineProps<{
 const emit = defineEmits<{
   select: [kind: 'video' | 'image']
   'video-engine': [engine: 'minimax' | 'ltx']
+  'video-start': [start: 'still' | 'text']
   'image-engine': [engine: 'flux' | 'krea']
   'music-engine': [engine: 'ace' | 'yue']
 }>()
@@ -149,6 +172,11 @@ function pickVideo(engine: 'minimax' | 'ltx') {
   if (props.current !== 'video') emit('select', 'video')
 }
 
+function pickStart(start: 'still' | 'text') {
+  emit('video-start', start)
+  if (props.current !== 'video') emit('select', 'video')
+}
+
 function pickImage(engine: 'flux' | 'krea') {
   emit('image-engine', engine)
   if (props.current !== 'image') emit('select', 'image')
diff --git a/pages/index.vue b/pages/index.vue
index 2fd2dcd..6c50b57 100644
--- a/pages/index.vue
+++ b/pages/index.vue
@@ -12,9 +12,11 @@
         :current="imageStudio ? 'image' : 'video'"
         :ltx-enabled="ltxEnabled"
         :video-engine="videoEngine"
+        :video-start="videoStart"
         :image-engine="v2Engine"
         @select="selectStudioKind"
         @video-engine="selectVideoEngine"
+        @video-start="selectVideoStart"
         @image-engine="setV2Engine"
       />
       <div class="grid min-w-0 gap-6 lg:grid-cols-[minmax(0,1fr)_minmax(0,1fr)] lg:items-stretch">
@@ -37,11 +39,13 @@
               : 'Edit: one still plus text. Compose is a separate graph if you need a second still.')
             : studioMode === 'edit'
             ? 'Drop the still to edit. An optional second still applies image 2 onto image 1. Files leave the desktop after save.'
-            : 'Drop a still, describe the motion, then send the job. Generate stays available — a long shot list takes one queue slot, and you can line up more while it runs.' }}</p>
+            : (textToVideo
+              ? 'Text-to-video: write a prompt and generate. No start still required. Shot 2+ still continues from the last frame.'
+              : 'Drop a still, describe the motion, then send the job. Switch the Video card to Text for text-to-video.') }}</p>
         </div>
 
         <div
-          v-if="!(studioMode === 'editv2' && (v2Mode === 'generate' || (v2Mode === 'iterate' && !file && !inputStillId)))"
+          v-if="!(studioMode === 'editv2' && (v2Mode === 'generate' || (v2Mode === 'iterate' && !file && !inputStillId))) && !(studioMode === 'video' && textToVideo && !file && !inputStillId)"
           class="relative cursor-pointer rounded-2xl border border-dashed border-white/15 bg-zinc-950/50 p-4 transition hover:border-amber-300/50"
           :class="{ 'border-amber-300/70 bg-amber-400/5': dragging }"
           @dragover.prevent="dragging = true"
@@ -73,6 +77,13 @@
             <p class="mt-1 text-sm text-zinc-500">{{ textToVideo ? 'Text-to-video does not need a still. Shot 2+ will still extend from the last frame.' : 'Browse the library or upload a new PNG, JPG, or WEBP' }}</p>
           </div>
         </div>
+        <div
+          v-else-if="studioMode === 'video' && textToVideo"
+          class="rounded-2xl border border-dashed border-white/15 bg-zinc-950/40 px-4 py-3 text-sm text-zinc-400"
+        >
+          <button type="button" class="font-medium text-amber-100 hover:underline" @click="openPicker('main')">Optional · add a start still</button>
+          <span class="mt-1 block text-zinc-500">Text-to-video does not need one. Shot 2+ still extends from the last frame.</span>
+        </div>
         <div
           v-else-if="studioMode === 'editv2' && v2Mode === 'generate' && (file || inputStillId)"
           class="rounded-2xl border border-white/10 bg-zinc-950/40 px-4 py-3 text-sm text-zinc-400"
@@ -649,30 +660,7 @@
         </div>
 
         <div v-if="studioMode === 'video'" class="space-y-3">
-          <div>
-            <p class="mb-2 text-xs uppercase tracking-wider text-zinc-500">Start from</p>
-            <div class="grid grid-cols-2 gap-2">
-              <button
-                type="button"
-                class="rounded-xl border px-3 py-2 text-left text-sm"
-                :class="videoStart === 'still' ? 'border-amber-300 bg-amber-400/10 text-amber-100' : 'border-white/10 hover:border-white/20'"
-                @click="videoStart = 'still'"
-              >
-                <span class="block font-semibold">A still</span>
-                <span class="text-xs text-zinc-400">Image to video. Extensions use the last frame.</span>
-              </button>
-              <button
-                type="button"
-                class="rounded-xl border px-3 py-2 text-left text-sm"
-                :class="videoStart === 'text' ? 'border-amber-300 bg-amber-400/10 text-amber-100' : 'border-white/10 hover:border-white/20'"
-                @click="videoStart = 'text'"
-              >
-                <span class="block font-semibold">Text only</span>
-                <span class="text-xs text-zinc-400">No start still. Shot 2+ still continues from last frame.</span>
-              </button>
-            </div>
-            <p class="mt-2 text-xs text-zinc-500">{{ videoModelHint }}</p>
-          </div>
+          <p class="text-xs text-zinc-500">{{ videoModelHint }}</p>
           <div v-if="showLegacyMinimaxV1 && videoEngine === 'minimax' && videoStart === 'still'">
             <p class="mb-2 text-xs uppercase tracking-wider text-zinc-500">MiniMax graph</p>
             <div class="grid grid-cols-2 gap-2">
@@ -7468,6 +7456,11 @@ function selectVideoEngine(engine: VideoEngineId) {
   applyEngineDefaults(engine)
 }
 
+function selectVideoStart(start: VideoStartId) {
+  videoStart.value = start
+  if (studioMode.value !== 'video') selectStudioKind('video')
+}
+
 function applyVideoWorkflow(raw?: string) {
   const parsed = coerceVideoWorkflow(raw, ltxEnabled.value)
   videoEngine.value = videoEngineOf(parsed)
@@ -7522,15 +7515,17 @@ async function generate() {
     toast('Choose a library folder before generating.')
     return
   }
-  try {
-    await ensureVideoInputStill()
-  } catch {
-    toast('Could not load the selected still for image-to-video.')
-    return
-  }
-  if (!textToVideo.value && !file.value && !inputStillId.value) {
-    toast('Load a still first. Image-to-video needs an input still.')
-    return
+  if (!textToVideo.value) {
+    try {
+      await ensureVideoInputStill()
+    } catch {
+      toast('Could not load the selected still for image-to-video.')
+      return
+    }
+    if (!file.value && !inputStillId.value) {
+      toast('Load a still first. Image-to-video needs an input still.')
+      return
+    }
   }
   const shots = shotScriptMode.value ? parsedShots.value : null
   const initialPrompt = (shots?.[0]?.prompt || prompt.value).trim()
@@ -7549,8 +7544,10 @@ async function generate() {
   try {
     await ensureGenerationNotifyPermission()
     const body = new FormData()
-    if (file.value) body.append('image', file.value)
-    if (inputStillId.value) body.append('stillId', inputStillId.value)
+    if (!textToVideo.value) {
+      if (file.value) body.append('image', file.value)
+      if (inputStillId.value) body.append('stillId', inputStillId.value)
+    }
     body.append('prompt', composePromptParts(promptPre.value, initialPrompt, promptPost.value) || initialPrompt)
     body.append('promptMid', initialPrompt)
     body.append('promptPre', promptPre.value.trim())