From a2f728fe2eeefceabd23544131343930278b9fbf Mon Sep 17 00:00:00 2001
From: Towsty <ianjmcarr@gmail.com>
Date: Fri, 28 Aug 2026 15:42:59 -0500
Subject: [PATCH] Flatten image edit controls and let CFG/steps be typed past
 the old slider caps.

Co-authored-by: Cursor <cursoragent@cursor.com>
---
 pages/index.vue                   | 176 ++++++++++++++++++------------
 server/api/edit.post.ts           |  17 +--
 server/api/extend.post.ts         |   7 +-
 server/api/generate.post.ts       |   6 +-
 server/utils/generationPresets.ts |  15 +--
 utils/generationPresets.ts        |  30 +++++
 6 files changed, 151 insertions(+), 100 deletions(-)

diff --git a/pages/index.vue b/pages/index.vue
index fec4ca1..aff8698 100644
--- a/pages/index.vue
+++ b/pages/index.vue
@@ -173,21 +173,37 @@
           <p v-else-if="identityPrompting" class="mb-2 text-xs text-zinc-500">
             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"
-            :rows="shotScriptMode || identityPrompting ? 12 : 6"
-            class="hidden max-h-[min(28rem,70vh)] w-full overflow-y-auto rounded-2xl border border-white/10 bg-zinc-950/70 px-4 py-3 text-sm outline-none ring-amber-300/40 placeholder:text-zinc-600 [overflow-wrap:anywhere] focus:ring-2 md:block"
-            :placeholder="promptPlaceholder"
-          />
-          <button
-            type="button"
-            class="w-full rounded-2xl border border-white/10 bg-zinc-950/70 px-4 py-3 text-left md:hidden"
-            @click="promptDrawerOpen = true"
+          <div
+            class="grid gap-3"
+            :class="studioMode === 'edit' ? 'md:grid-cols-2' : ''"
           >
-            <p v-if="prompt.trim()" class="line-clamp-4 whitespace-pre-wrap text-sm text-zinc-200 [overflow-wrap:anywhere]">{{ prompt }}</p>
-            <p v-else class="text-sm text-zinc-600">{{ promptPlaceholder }}</p>
-            <span class="mt-2 inline-block text-xs text-amber-200">{{ prompt.trim() ? 'View / edit prompt' : 'Write prompt' }}</span>
-          </button>
+            <div class="min-w-0">
+              <textarea
+                v-model="prompt"
+                :rows="shotScriptMode || identityPrompting ? 12 : 6"
+                class="hidden max-h-[min(28rem,70vh)] w-full overflow-y-auto rounded-2xl border border-white/10 bg-zinc-950/70 px-4 py-3 text-sm outline-none ring-amber-300/40 placeholder:text-zinc-600 [overflow-wrap:anywhere] focus:ring-2 md:block"
+                :placeholder="promptPlaceholder"
+              />
+              <button
+                type="button"
+                class="w-full rounded-2xl border border-white/10 bg-zinc-950/70 px-4 py-3 text-left md:hidden"
+                @click="promptDrawerOpen = true"
+              >
+                <p v-if="prompt.trim()" class="line-clamp-4 whitespace-pre-wrap text-sm text-zinc-200 [overflow-wrap:anywhere]">{{ prompt }}</p>
+                <p v-else class="text-sm text-zinc-600">{{ promptPlaceholder }}</p>
+                <span class="mt-2 inline-block text-xs text-amber-200">{{ prompt.trim() ? 'View / edit prompt' : 'Write prompt' }}</span>
+              </button>
+            </div>
+            <label v-if="studioMode === 'edit'" class="block min-w-0 text-sm">
+              <span class="mb-2 block font-medium text-zinc-300">Negative prompt</span>
+              <textarea
+                v-model="editNegative"
+                rows="2"
+                class="min-h-[4.5rem] w-full rounded-2xl border border-white/10 bg-zinc-950/70 px-4 py-3 text-sm outline-none ring-amber-300/40 placeholder:text-zinc-600 [overflow-wrap:anywhere] focus:ring-2 md:h-[calc(100%-1.75rem)] md:min-h-[8.5rem]"
+                placeholder="What to avoid…"
+              />
+            </label>
+          </div>
           <p v-if="shotScriptMode && studioMode === 'video' && parsedShots.length" class="mt-2 text-xs text-zinc-500 md:hidden">
             {{ parsedShots.length }} shot{{ parsedShots.length === 1 ? '' : 's' }} in this script. Open the prompt to scroll through them.
           </p>
@@ -601,16 +617,50 @@
             @remove="deleteGenerationPreset"
           />
         </div>
-        <div v-else class="max-w-md space-y-2">
-          <LoraSelect v-model="imageLoraStack" :names="imageLoraOptions" label="LoRA" empty-label="None" compact />
-          <GenerationPresetBar
-            :presets="imageGenerationPresets"
-            v-model:selected-id="imagePresetId"
-            v-model:name="imagePresetName"
-            @load="loadGenerationPreset"
-            @save="saveGenerationPreset"
-            @remove="deleteGenerationPreset"
-          />
+        <div v-else class="space-y-3">
+          <p class="text-xs text-zinc-500">Flux.2 Klein 9B. Shares the desktop GPU with video.</p>
+          <div class="max-w-md space-y-2">
+            <LoraSelect v-model="imageLoraStack" :names="imageLoraOptions" label="LoRA" empty-label="None" compact />
+            <GenerationPresetBar
+              :presets="imageGenerationPresets"
+              v-model:selected-id="imagePresetId"
+              v-model:name="imagePresetName"
+              @load="loadGenerationPreset"
+              @save="saveGenerationPreset"
+              @remove="deleteGenerationPreset"
+            />
+          </div>
+          <div class="grid grid-cols-2 gap-3">
+            <label class="block text-sm">
+              <span class="mb-1 block text-zinc-400">Steps</span>
+              <input
+                v-model.number="editSteps"
+                type="number"
+                :min="IMAGE_STEPS_MIN"
+                :max="IMAGE_STEPS_MAX"
+                step="1"
+                class="w-full rounded-xl border border-white/10 bg-zinc-950 px-3 py-2 text-sm outline-none ring-amber-300/40 focus:ring-2"
+                @change="onEditStepsNumber"
+              >
+            </label>
+            <label class="block text-sm">
+              <span class="mb-1 block text-zinc-400">CFG</span>
+              <input
+                v-model.number="editCfg"
+                type="number"
+                :min="IMAGE_CFG_MIN"
+                :max="IMAGE_CFG_MAX"
+                :step="IMAGE_CFG_STEP"
+                class="w-full rounded-xl border border-white/10 bg-zinc-950 px-3 py-2 text-sm outline-none ring-amber-300/40 focus:ring-2"
+                @change="onEditCfgNumber"
+              >
+            </label>
+            <label class="col-span-2 block text-sm">
+              <span class="mb-1 block text-zinc-400">Seed</span>
+              <input v-model="seedInput" class="w-full rounded-xl border border-white/10 bg-zinc-950 px-3 py-2 text-sm" placeholder="random">
+            </label>
+          </div>
+          <p class="text-xs text-zinc-500">Graph default is 20 steps, CFG 1, euler. CFG 0–20 (0.1 steps), steps 1–50. A second still switches the two-image branch on.</p>
         </div>
 
         <details v-if="studioMode === 'video'" class="rounded-2xl border border-white/10 bg-zinc-950/40 px-4 py-3">
@@ -660,11 +710,11 @@
                   :min="CFG_MIN"
                   :max="CFG_MAX"
                   :step="CFG_STEP"
-                  class="w-20 rounded-xl border border-white/10 bg-zinc-950 px-3 py-2 text-sm outline-none ring-amber-300/40 focus:ring-2"
+                  class="w-24 rounded-xl border border-white/10 bg-zinc-950 px-3 py-2 text-sm outline-none ring-amber-300/40 focus:ring-2"
                   @change="onCfgNumber"
                 >
               </div>
-              <p class="mt-1 text-[11px] text-zinc-500">{{ ltxVideo ? 'Default 3.5 for PinkCherry' : (turbo ? 'Default 1.5 with Turbo LoRA' : 'Default 4.0 with High-Fidelity') }}</p>
+              <p class="mt-1 text-[11px] text-zinc-500">{{ ltxVideo ? 'Default 3.5 for PinkCherry' : (turbo ? 'Default 1.5 with Turbo LoRA' : 'Default 4.0 with High-Fidelity') }}. Type 0–20 (0.1 steps); the box is the source of truth.</p>
             </div>
             <div class="grid grid-cols-2 gap-3">
               <label class="text-sm">
@@ -688,7 +738,6 @@
               <span class="mb-1 block text-zinc-400">Seed</span>
               <input v-model="seedInput" class="w-full rounded-xl border border-white/10 bg-zinc-950 px-3 py-2" placeholder="random">
             </label>
-            <LoraSelect v-model="videoLoraStack" :names="videoLoraOptions" label="LoRA" empty-label="None" :clip-strength="false" />
             <div>
               <p class="mb-2 text-xs uppercase tracking-wider text-zinc-500">Sampler & scheduler</p>
               <div class="grid grid-cols-2 gap-3">
@@ -713,35 +762,6 @@
           </div>
         </details>
 
-        <details v-else class="rounded-2xl border border-white/10 bg-zinc-950/40 px-4 py-3">
-          <summary class="cursor-pointer text-sm font-medium">Image edit settings</summary>
-          <div class="mt-4 space-y-4">
-            <p class="text-xs text-zinc-500">
-              Flux.2 Klein 9B. Shares the desktop GPU with video.
-            </p>
-            <label class="block text-sm">
-              <span class="mb-1 block text-zinc-400">Negative prompt</span>
-              <textarea v-model="editNegative" rows="2" class="w-full rounded-xl border border-white/10 bg-zinc-950 px-3 py-2 text-sm" />
-            </label>
-            <div class="grid gap-4 sm:grid-cols-2">
-              <label class="block text-sm">
-                <span class="mb-1 block text-zinc-400">Steps · {{ editSteps }}</span>
-                <input v-model.number="editSteps" type="range" min="4" max="40" step="1" class="w-full">
-              </label>
-              <label class="block text-sm">
-                <span class="mb-1 block text-zinc-400">CFG · {{ editCfg }}</span>
-                <input v-model.number="editCfg" type="range" min="0.5" max="4" step="0.1" class="w-full">
-              </label>
-              <label class="block text-sm sm:col-span-2">
-                <span class="mb-1 block text-zinc-400">Seed</span>
-                <input v-model="seedInput" class="w-full rounded-xl border border-white/10 bg-zinc-950 px-3 py-2 text-sm" placeholder="random">
-              </label>
-            </div>
-            <LoraSelect v-model="imageLoraStack" :names="imageLoraOptions" label="LoRA" empty-label="None" />
-            <p class="text-xs text-zinc-500">Graph default is 20 steps, CFG 1, euler. A second still switches the two-image branch on.</p>
-          </div>
-        </details>
-
         <label
           v-if="studioMode === 'video' && queuedExtensionCount > 0"
           class="flex cursor-pointer items-start gap-3 rounded-2xl border border-white/10 bg-zinc-950/40 px-4 py-3"
@@ -1605,7 +1625,20 @@ import {
   readLoraStack,
   type LoraStackItem
 } from '~/utils/loras'
-import type { GenerationPreset } from '~/utils/generationPresets'
+import {
+  clampImageCfg,
+  clampImageSteps,
+  clampVideoCfg,
+  IMAGE_CFG_MAX,
+  IMAGE_CFG_MIN,
+  IMAGE_CFG_STEP,
+  IMAGE_STEPS_MAX,
+  IMAGE_STEPS_MIN,
+  VIDEO_CFG_MAX,
+  VIDEO_CFG_MIN,
+  VIDEO_CFG_STEP,
+  type GenerationPreset
+} from '~/utils/generationPresets'
 import {
   coerceVideoWorkflow,
   composeVideoWorkflow,
@@ -1800,9 +1833,9 @@ const videoPresetId = ref('')
 const imagePresetId = ref('')
 const videoPresetName = ref('')
 const imagePresetName = ref('')
-const CFG_MIN = 1
-const CFG_MAX = 10
-const CFG_STEP = 0.5
+const CFG_MIN = VIDEO_CFG_MIN
+const CFG_MAX = VIDEO_CFG_MAX
+const CFG_STEP = VIDEO_CFG_STEP
 const CFG_TURBO = 1.5
 const CFG_QUALITY = 4
 const cfg = ref(CFG_TURBO)
@@ -2847,8 +2880,15 @@ function clampDuration(value: number) {
 }
 
 function clampCfg(value: number) {
-  if (!Number.isFinite(value)) return turbo.value ? CFG_TURBO : CFG_QUALITY
-  return Math.min(CFG_MAX, Math.max(CFG_MIN, Math.round(value / CFG_STEP) * CFG_STEP))
+  return clampVideoCfg(value, turbo.value ? CFG_TURBO : CFG_QUALITY)
+}
+
+function onEditStepsNumber() {
+  editSteps.value = clampImageSteps(editSteps.value)
+}
+
+function onEditCfgNumber() {
+  editCfg.value = clampImageCfg(editCfg.value)
 }
 
 function onCfgSlider(event: Event) {
@@ -2925,8 +2965,8 @@ function currentPresetSnapshot() {
       kind: 'image' as const,
       loraStack: filterLoraStackForStudio(imageLoraStack.value, ltxEnabled.value),
       settings: {
-        steps: editSteps.value,
-        cfg: editCfg.value,
+        steps: clampImageSteps(editSteps.value),
+        cfg: clampImageCfg(editCfg.value),
         negative: editNegative.value
       }
     }
@@ -2954,8 +2994,8 @@ function applyGenerationPreset(preset: GenerationPreset) {
   if (preset.kind === 'image') {
     imageLoraStack.value = stack
     const settings = preset.settings as { steps?: number; cfg?: number; negative?: string }
-    if (typeof settings.steps === 'number') editSteps.value = settings.steps
-    if (typeof settings.cfg === 'number') editCfg.value = settings.cfg
+    if (typeof settings.steps === 'number') editSteps.value = clampImageSteps(settings.steps)
+    if (typeof settings.cfg === 'number') editCfg.value = clampImageCfg(settings.cfg)
     if (typeof settings.negative === 'string') editNegative.value = settings.negative
     imagePresetId.value = preset.id
     imagePresetName.value = preset.name
@@ -4603,8 +4643,8 @@ async function editImage() {
     if (editRefFile.value) body.append('image2', editRefFile.value)
     body.append('prompt', prompt.value.trim())
     body.append('negative', editNegative.value)
-    body.append('steps', String(editSteps.value))
-    body.append('cfg', String(editCfg.value))
+    body.append('steps', String(clampImageSteps(editSteps.value)))
+    body.append('cfg', String(clampImageCfg(editCfg.value)))
     body.append('seed', seedInput.value || 'random')
     body.append('folderId', folderId.value)
     body.append('hideInput', String(showPrivacyToggles.value && hideInputPreview.value))
diff --git a/server/api/edit.post.ts b/server/api/edit.post.ts
index c210523..30af520 100644
--- a/server/api/edit.post.ts
+++ b/server/api/edit.post.ts
@@ -2,18 +2,7 @@ import { addStudioJob, kickStudioQueue, listStudioJobs } from '~/server/utils/st
 import { parsePostedLoraStack, listStudioLoras, persistLoraFields } from '~/server/utils/loras'
 import { comfyConfigured } from '~/server/utils/comfy'
 import { imageDimensions } from '~/server/utils/resolution'
-
-function parseSteps(raw: unknown) {
-  const value = Number(raw)
-  if (!Number.isFinite(value)) return 20
-  return Math.min(40, Math.max(1, Math.round(value)))
-}
-
-function parseCfg(raw: unknown) {
-  const value = Number(raw)
-  if (!Number.isFinite(value)) return 1
-  return Math.min(8, Math.max(0.5, Math.round(value * 10) / 10))
-}
+import { clampImageCfg, clampImageSteps } from '~/utils/generationPresets'
 
 function parsePasses(raw: string | undefined) {
   if (!raw) return [] as { prompt: string }[]
@@ -75,8 +64,8 @@ export default defineEventHandler(async (event) => {
   const folderLocked = Boolean(destFolder?.protected && !destFolder.unlocked)
   const hideInput = fields.hideInput === 'true'
   const hideThumbnail = fields.hideThumbnail === 'true'
-  const steps = parseSteps(fields.steps)
-  const cfg = parseCfg(fields.cfg)
+  const steps = clampImageSteps(fields.steps)
+  const cfg = clampImageCfg(fields.cfg)
   const seed = fields.seed && fields.seed !== 'random'
     ? Number(fields.seed)
     : Math.floor(Math.random() * 2_147_483_647)
diff --git a/server/api/extend.post.ts b/server/api/extend.post.ts
index a94df07..0f4c2d9 100644
--- a/server/api/extend.post.ts
+++ b/server/api/extend.post.ts
@@ -3,6 +3,7 @@ import { join } from 'node:path'
 import { composeShotPrompt } from '~/utils/globalLocks'
 import { parsePostedLoraStack, listStudioLoras, persistLoraFields } from '~/server/utils/loras'
 import { readLoraStack } from '~/utils/loras'
+import { clampVideoCfg } from '~/utils/generationPresets'
 import { defaultVideoSteps, isLtxWorkflow, LTX_DISABLED_MESSAGE, ltxWorkflowEnabled, parseVideoWorkflow } from '~/utils/videoModels'
 
 const SAMPLERS = new Set(['res_multistep', 'euler', 'dpmpp_2m'])
@@ -24,11 +25,7 @@ function parseFps(raw: unknown) {
 }
 
 function parseCfg(raw: unknown, turbo: boolean) {
-  const fallback = turbo ? 1.5 : 4
-  const value = Number(raw)
-  if (!Number.isFinite(value)) return fallback
-  const snapped = Math.round(value * 2) / 2
-  return Math.min(10, Math.max(1, snapped))
+  return clampVideoCfg(raw, turbo ? 1.5 : 4)
 }
 
 function parseSampler(raw: unknown) {
diff --git a/server/api/generate.post.ts b/server/api/generate.post.ts
index 91ea93a..a43f8cd 100644
--- a/server/api/generate.post.ts
+++ b/server/api/generate.post.ts
@@ -1,6 +1,7 @@
 import { addStudioJob, kickStudioQueue, listStudioJobs } from '~/server/utils/studioQueue'
 import { listStudioLoras, parsePostedLoraStack, parseShotLoraStacks, persistLoraFields } from '~/server/utils/loras'
 import type { LoraStackItem } from '~/utils/loras'
+import { clampVideoCfg } from '~/utils/generationPresets'
 import { defaultVideoSteps, isLtxWorkflow, isTextToVideo, LTX_DISABLED_MESSAGE, ltxWorkflowEnabled, parseVideoWorkflow } from '~/utils/videoModels'
 import { allowIdentityRefs, normalizePermanenceRefs, resolveGlobalLocks, type PermanenceRef } from '~/utils/globalLocks'
 
@@ -66,10 +67,7 @@ function parseFps(raw: string | undefined) {
 
 function parseCfg(raw: string | undefined, turbo: boolean, ltx = false) {
   const fallback = ltx ? 3.5 : (turbo ? 1.5 : 4)
-  const value = Number(raw)
-  if (!Number.isFinite(value)) return fallback
-  const snapped = Math.round(value * 2) / 2
-  return Math.min(10, Math.max(1, snapped))
+  return clampVideoCfg(raw, fallback)
 }
 
 function parseSampler(raw: string | undefined, ltx = false) {
diff --git a/server/utils/generationPresets.ts b/server/utils/generationPresets.ts
index 4dddce5..d30d50d 100644
--- a/server/utils/generationPresets.ts
+++ b/server/utils/generationPresets.ts
@@ -3,6 +3,9 @@ import { join } from 'node:path'
 import { ownerRoot } from '~/server/utils/library'
 import { filterLoraStackForStudio, normalizeLoraStack } from '~/utils/loras'
 import {
+  clampImageCfg,
+  clampImageSteps,
+  clampVideoCfg,
   normalizePresetName,
   parseGenerationPresetKind,
   type GenerationPreset,
@@ -54,12 +57,6 @@ function mutate<T>(owner: string, fn: (presets: GenerationPreset[]) => T): Promi
   return run
 }
 
-function clampInt(raw: unknown, fallback: number, min: number, max: number) {
-  const value = Number(raw)
-  if (!Number.isFinite(value)) return fallback
-  return Math.min(max, Math.max(min, Math.round(value)))
-}
-
 function clampNum(raw: unknown, fallback: number, min: number, max: number, step = 0.1) {
   const value = Number(raw)
   if (!Number.isFinite(value)) return fallback
@@ -77,7 +74,7 @@ function sanitizeVideoSettings(raw: unknown): VideoPresetSettings {
   return {
     workflow,
     turbo,
-    cfg: clampNum(rec.cfg, turbo ? 1.5 : 4, 1, 10, 0.5),
+    cfg: clampVideoCfg(rec.cfg, turbo ? 1.5 : 4),
     fps: rec.fps === 12 || rec.fps === 30 || rec.fps === 24 ? Number(rec.fps) : 24,
     samplerName: sampler === 'euler' || sampler === 'dpmpp_2m' || sampler === 'res_multistep' ? sampler : 'res_multistep',
     scheduler: scheduler === 'ddim_uniform' || scheduler === 'sgm_uniform' || scheduler === 'simple' ? scheduler : 'simple',
@@ -90,8 +87,8 @@ function sanitizeVideoSettings(raw: unknown): VideoPresetSettings {
 function sanitizeImageSettings(raw: unknown): ImagePresetSettings {
   const rec = raw && typeof raw === 'object' ? raw as Record<string, unknown> : {}
   return {
-    steps: clampInt(rec.steps, 20, 1, 40),
-    cfg: clampNum(rec.cfg, 1, 0.5, 4, 0.1),
+    steps: clampImageSteps(rec.steps),
+    cfg: clampImageCfg(rec.cfg),
     negative: String(rec.negative || '').slice(0, 2000)
   }
 }
diff --git a/utils/generationPresets.ts b/utils/generationPresets.ts
index c2923d7..6acdfb7 100644
--- a/utils/generationPresets.ts
+++ b/utils/generationPresets.ts
@@ -21,6 +21,36 @@ export type ImagePresetSettings = {
   negative?: string
 }
 
+export const IMAGE_STEPS_MIN = 1
+export const IMAGE_STEPS_MAX = 50
+export const IMAGE_CFG_MIN = 0
+export const IMAGE_CFG_MAX = 20
+export const IMAGE_CFG_STEP = 0.1
+
+export const VIDEO_CFG_MIN = 0
+export const VIDEO_CFG_MAX = 20
+export const VIDEO_CFG_STEP = 0.1
+
+export function clampImageSteps(raw: unknown, fallback = 20) {
+  const value = Number(raw)
+  if (!Number.isFinite(value)) return fallback
+  return Math.min(IMAGE_STEPS_MAX, Math.max(IMAGE_STEPS_MIN, Math.round(value)))
+}
+
+export function clampImageCfg(raw: unknown, fallback = 1) {
+  const value = Number(raw)
+  if (!Number.isFinite(value)) return fallback
+  const snapped = Math.round(value / IMAGE_CFG_STEP) * IMAGE_CFG_STEP
+  return Math.min(IMAGE_CFG_MAX, Math.max(IMAGE_CFG_MIN, Math.round(snapped * 10) / 10))
+}
+
+export function clampVideoCfg(raw: unknown, fallback = 1.5) {
+  const value = Number(raw)
+  if (!Number.isFinite(value)) return fallback
+  const snapped = Math.round(value / VIDEO_CFG_STEP) * VIDEO_CFG_STEP
+  return Math.min(VIDEO_CFG_MAX, Math.max(VIDEO_CFG_MIN, Math.round(snapped * 10) / 10))
+}
+
 export type GenerationPreset = {
   id: string
   name: string