towsty/aigen
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

Send V2 identity prompts to Reference to Video instead of Image to Video.

Browse Source 
Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
Towsty
2026-08-26 23:08:43 -05:00
co-authored by Cursor
parent e334cb7a7a
commit d280985553
3 changed files with 33 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files
pages/index.vue
+14 -8
View File
@@ -109,10 +109,10 @@
</div>
<p v-if="shotScriptMode && studioMode === 'video'" class="mb-2 text-xs text-zinc-500">
Paste a block. Text before the first marker is shot 1. Put <span class="font-medium text-zinc-300">shot 2</span>, <span class="font-medium text-zinc-300">shot 3</span> on their own lines to queue extensions — a line like <span class="font-medium text-zinc-300">Shot Type</span> will not start a new shot. Every shot uses the duration slider.
<span v-if="identityPrompting"> Identity boilerplate is prepended to the first shot only.</span>
<span v-if="identityPrompting"> Identity boilerplate is prepended to the first shot only, then concatenated into Reference to Video.</span>
</p>
<p v-else-if="identityPrompting" class="mb-2 text-xs text-zinc-500">
Write camera, action, and audio. We prepend the <span class="font-medium text-zinc-300">&lt;Picture 1&gt;</span> identity lock automatically.
Write camera, action, and audio. That is the dynamic action prompt. We concatenate the <span class="font-medium text-zinc-300">&lt;Picture 1&gt;</span> identity prefix in front of it and send the result to <span class="font-medium text-zinc-300">Reference to Video</span>, not Image to Video.
</p>
<textarea
v-model="prompt"
@@ -121,8 +121,13 @@
:placeholder="promptPlaceholder"
/>
<details v-if="identityPrompting && prompt.trim()" class="mt-2 rounded-xl border border-white/10 bg-zinc-950/40 px-3 py-2">
<summary class="cursor-pointer text-xs text-zinc-400">Prompt sent to MiniMax</summary>
<p class="mt-2 whitespace-pre-wrap text-xs text-zinc-300">{{ composedIdentityPrompt }}</p>
<summary class="cursor-pointer text-xs text-zinc-400">Prompt sent to Reference to Video</summary>
<p class="mt-2 text-[10px] uppercase tracking-wide text-zinc-500">Identity prefix</p>
<p class="mt-1 whitespace-pre-wrap text-xs text-zinc-400">{{ identityBoilerplate(extraIdentityPictures) }}</p>
<p class="mt-3 text-[10px] uppercase tracking-wide text-zinc-500">Dynamic action prompt</p>
<p class="mt-1 whitespace-pre-wrap text-xs text-zinc-300">{{ identityActionPrompt }}</p>
<p class="mt-3 text-[10px] uppercase tracking-wide text-zinc-500">Concatenated result</p>
<p class="mt-1 whitespace-pre-wrap text-xs text-zinc-300">{{ composedIdentityPrompt }}</p>
</details>
<div v-if="shotScriptMode && studioMode === 'video'" class="mt-3 space-y-2">
<p v-if="!parsedShots.length" class="text-xs text-zinc-500">Add a first shot, then <span class="text-zinc-300">shot 2</span> for an extension.</p>
@@ -1296,11 +1301,12 @@ const generateLabel = computed(() => {
const parsedShots = computed(() => parseShotScript(prompt.value))
const identityPrompting = computed(() => studioMode.value === 'video' && videoWorkflow.value === 'v2' && useIdentityRefs.value)
const extraIdentityPictures = computed(() => identityRefs.value.flatMap((item, index) => item ? [index + 2] : []))
const identityActionPrompt = computed(() => {
if (shotScriptMode.value) return parsedShots.value[0]?.prompt || prompt.value
return prompt.value
})
const composedIdentityPrompt = computed(() => {
const action = shotScriptMode.value
? (parsedShots.value[0]?.prompt || prompt.value)
: prompt.value
return buildIdentityPrompt(action, extraIdentityPictures.value)
return buildIdentityPrompt(identityActionPrompt.value, extraIdentityPictures.value)
})
const promptPlaceholder = computed(() => {
if (studioMode.value === 'edit') {