docs: add v0.32.0 changelog and publish CMS release notes

.cursor/skills/cms-changelog-sync/SKILL.md

@@ -91,12 +91,12 @@ Config: `.github/scripts/cms/cms-config.json` → `simplify`
 |------|-------|
 | Total bullets per version | **up to 10** (`max_bullets_total: 10`) |
 | Section headings max | **3** (`max_sections: 3`) |
-| Section order | **New Open-Source Model Support** → **New Node Updates** → **Partner Node Updates** |
+| Section order | **New Open-Source Model Support** → **Partner Node Updates** → **New Node Updates** (optional) |
 | Words per version | ~60–120 |
 | Bullet format | `[**Name**](pr_url): 6–12 words with one key trait` |
 | PR links | **Keep** when source has them; never invent URLs |
-| New Node Updates | Include meaningful entries from source **New Nodes** section (workflows, sockets, multimodal nodes) |
-| Drop | Bug fixes, performance, pure Load3D plumbing, internal refactors, **ComfyUI-WIKI dependency bumps** (see below) |
+| New Node Updates | **Optional by default.** Omit from CMS popup even if docs has **New Nodes**; add only when a human explicitly asks |
+| Drop | Bug fixes, performance, pure Load3D plumbing, internal refactors, **ComfyUI-WIKI dependency bumps** (see below), and New Nodes unless requested |
 
 Style: principle-only prompt in `cms-simplify-prompt.ts` (no concrete version examples — avoids LLM contamination).
 
@@ -112,19 +112,23 @@ When curating `changelog/index.mdx` from ComfyUI git history, **do not add bulle
 
 Also omit standalone **frontend package semver bumps** unless tied to a user-visible fix worth its own bullet. CMS simplify must never promote WIKI-only items into popup copy even if they appear in the full docs block.
 
-Example staging shape (placeholders only):
+Example staging shape (placeholders only). **New Node Updates** is optional and usually omitted:
 
 ```markdown
 **New Open-Source Model Support**
 * [**Model Name**](source_url): Short description with 1–2 traits from the release data
 
-**New Node Updates**
-* [**Node Name**](source_url): What the node does and why it matters
-
 **Partner Node Updates**
 * [**Partner Node**](source_url): Partner scope and capability from the release data
 ```
 
+Only when a human asks to include nodes:
+
+```markdown
+**New Node Updates**
+* [**Node Name**](source_url): What the node does and why it matters
+```
+
 Sync adds header: `# ComfyUI vX.Y.Z` via `format-cms-content.ts`.
 
 ## Projects (comfyui + cloud)
@@ -138,7 +142,7 @@ Same changelog content; Strapi `project` field and CMS header differ (`# ComfyUI
 | `comfyui` | `staging/{locale}/…` | `# ComfyUI vX.Y.Z` |
 | `cloud` | `staging/cloud/{locale}/…` | `# Cloud vX.Y.Z` |
 
-Prepare runs LLM once on comfyui, then **copies staging to cloud**. Sync/publish must be project-scoped by agents: `--project comfyui` first, then `--project cloud` only after explicit cloud approval.
+When prepare targets both projects, it runs the LLM once on comfyui, then **merges only the prepared version blocks into cloud** (does not overwrite older cloud-only tracking shortlinks). With `--project cloud` alone, cloud is prepared directly. Sync/publish must be project-scoped by agents: `--project comfyui` first, then `--project cloud` only after explicit cloud approval.
 
 Single project: `--project comfyui`, `--project cloud`, or `CMS_PROJECT=<project>`.
 
@@ -292,5 +296,6 @@ When user asks to update CMS release notes:
 |--|-----------|-----------|
 | Source | `changelog/index.mdx` | `staging/en/…` |
 | Length | Full detail | 3–5 bullets |
+| New Nodes | Keep in full changelog | **Optional**; omit by default unless a human asks |
 | i18n | `zh/changelog/` etc. | `staging/zh/` etc. |
 | Deploy | Mintlify | Strapi draft → publish |

.github/scripts/cms/README.md

@@ -60,6 +60,8 @@ Strapi drafts → publish → published-versions.json
 
 Each version is **saved immediately** after simplify/translate (safe to resume).
 
+After prepare, secondary projects (cloud) **merge only the prepared `<Update>` blocks** from comfyui. Older cloud blocks (including project-specific tracking shortlinks) are left untouched. Do not full-file overwrite `staging/cloud/`.
+
 ## Projects & attention
 
 | Project | CLI | Default attention |
@@ -98,13 +100,13 @@ Configured in `cms-config.json` → `simplify`:
 | Key | Default | Meaning |
 |-----|---------|---------|
 | `max_bullets_total` | **10** | Bullets for the **entire version** (not per section) |
-| `max_sections` | **3** | `**New Open-Source Model Support**` → `**New Node Updates**` → `**Partner Node Updates**` (bold labels, not `##`) |
+| `max_sections` | **3** | `**New Open-Source Model Support**` → `**Partner Node Updates**` → optional `**New Node Updates**` (bold labels, not `##`) |
 
-- Section order is **fixed**: open-source models first, node updates second, partner nodes last
-- Include **all meaningful New Nodes** entries from the docs changelog (workflows, output sockets, multimodal nodes)
+- Section order is **fixed** when present: open-source models first, partner nodes second, node updates last
+- **New Node Updates is optional by default.** Omit from the CMS popup even if docs has New Nodes; include only when a human explicitly asks
 - Each bullet: **[**Name**](pr_url): 12–25 word description** — preserve model/node traits from source
 - **Keep PR links** when the source has them
-- **Drop** performance tweaks, minor fixes, Load3D/UI housekeeping
+- **Drop** performance tweaks, minor fixes, Load3D/UI housekeeping, and New Nodes unless requested
 - English only (Step 1): `pnpm cms:prepare:en -- --force v0.25.0`
 - Translate only (Step 2): `pnpm cms:prepare:locales -- --force v0.25.0` — reads existing `staging/en/`, never re-simplifies
 

.github/scripts/cms/cms-simplify-prompt.ts

@@ -24,26 +24,25 @@ export const CMS_SIMPLIFY_SYSTEM_PROMPT = `You are an expert technical writer cr
 
 **Section order (mandatory — never reorder):**
 1. \`**New Open-Source Model Support**\` — when the source lists new open-source models
-2. \`**New Node Updates**\` — when the source has a **New Nodes** (or equivalent) section with user-facing additions
-3. \`**Partner Node Updates**\` — when the source lists partner or API node updates; always last among these three
+2. \`**Partner Node Updates**\` — when the source lists partner or API node updates; when open-source models are also present, place this immediately after them
+3. \`**New Node Updates**\` — **optional**; omit by default even if the source has a **New Nodes** section
 
-Never place Partner Node Updates before New Open-Source Model Support. Never merge categories into a flat list. Omit a section entirely if the source has no items for it.
+Emit only sections that have source items, in the order above. When open-source models are absent, Partner Node Updates may lead. Never place Partner Node Updates before New Open-Source Model Support when both are present. Never place New Node Updates before Partner Node Updates when both are present. Never merge categories into a flat list. Omit a section entirely if the source has no items for it.
 
 **New Open-Source Model Support:**
 - Include every open-source model from the source (within the bullet limit)
-- Do not drop models to make room for nodes or partner items
+- Do not drop models to make room for partner items
 - Preserve 1–2 distinguishing traits per model from the source (variants, sizes, encoder, modality, key capability)
 
-**New Node Updates:**
-- When the source has a **New Nodes** section, include meaningful user-facing entries (within the bullet limit)
-- Include: new workflow nodes, save/load or I/O improvements, multimodal or video node additions
-- Skip only entries that are minor plumbing with no user workflow impact (e.g. internal loader tweaks)
-- Use \`**New Node Updates**\` as the section label
-
 **Partner Node Updates:**
 - Include partner/API node additions or updates from the source
 - Preserve scope or capability details when stated (e.g. number of new nodes, supported modality)
-- Always in the last section when other sections are present
+- Always the second section when open-source models are also present (right after Open-Source Model Support)
+
+**New Node Updates (optional — default omit):**
+- **Default: do not emit this section.** Docs changelog may list New Nodes; the CMS popup does not need them.
+- Only include \`**New Node Updates**\` when the user message explicitly asks to include new nodes (or equivalent). Otherwise skip the whole section even if the source has **New Nodes**.
+- When explicitly requested: include meaningful user-facing entries (within the bullet limit); last among the three sections; skip minor plumbing with no workflow impact
 
 **Bullet format:**
 - Linked: * [**Name**](url_from_source): Description
@@ -56,10 +55,11 @@ Never place Partner Node Updates before New Open-Source Model Support. Never mer
 
 **What to include (priority — matches section order):**
 1. All open-source models from the source
-2. All meaningful **New Nodes** entries
-3. Partner/API node updates
+2. Partner/API node updates
+3. **New Nodes only if the user message explicitly requests them** (otherwise omit)
 
 **What to drop:**
+- **New Node Updates / New Nodes** by default (unless explicitly requested in the user message)
 - Minor fixes, refactors, dtype cleanups, internal tooling
 - Pure loader/plumbing changes with no workflow impact
 - Performance, stability, API housekeeping, console logging
@@ -81,7 +81,8 @@ export function buildSimplifyUserPrompt(
     "Rewrite the release note below for an **in-app popup**.",
     "",
     `Hard limits: **${limits.maxBulletsTotal} bullets total**, **${limits.maxSections} section headings max**.`,
-    "Section order: New Open-Source Model Support → New Node Updates → Partner Node Updates (omit empty sections).",
+    "Section order: New Open-Source Model Support → Partner Node Updates → (optional) New Node Updates.",
+    "Omit **New Node Updates** by default even if the source has New Nodes, unless this message explicitly asks to include them.",
     "Section labels: **bold** (e.g. **Partner Node Updates**), not ## headings.",
     "Use only facts and links from the release data. Each bullet: **6–12 words** — one key trait, no filler.",
     "",

.github/scripts/cms/cms-staging-io.ts

@@ -2,14 +2,14 @@
  * Shared read/write helpers for CMS staging MDX (per-version checkpoint).
  */
 
-import { copyFile, mkdir, readFile, writeFile } from "fs/promises";
+import { mkdir, readFile, writeFile } from "fs/promises";
 import { dirname, join } from "path";
 import {
   parseDocument,
   serializeUpdateBlocksDocument,
   sortUpdateBlocksByVersion,
 } from "../i18n/chunked-translate.ts";
-import { parseChangelogUpdates } from "./changelog-parse.ts";
+import { normalizeVersion, parseChangelogUpdates } from "./changelog-parse.ts";
 import { configForProject, type CmsConfig } from "./cms-config.ts";
 import { isEnoent, ROOT } from "./cms-env.ts";
 
@@ -66,12 +66,26 @@ export async function writeStagingCheckpoint(
   return output;
 }
 
-/** Copy all locale staging files from one project to another (same changelog body). */
+/**
+ * Merge selected version blocks from one project's staging into another.
+ *
+ * Never replaces whole locale files: cloud may keep project-specific tracking
+ * shortlinks on older versions (e.g. clo* vs rel*). Only upsert the versions
+ * prepared in this run.
+ */
 export async function copyProjectStaging(
   baseConfig: CmsConfig,
   fromProject: string,
-  toProject: string
+  toProject: string,
+  versions: string[]
 ): Promise<number> {
+  const versionList = [
+    ...new Set(versions.map((v) => normalizeVersion(v)).filter(Boolean)),
+  ];
+  if (versionList.length === 0) {
+    return 0;
+  }
+
   const fromConfig = configForProject(baseConfig, fromProject);
   const toConfig = configForProject(baseConfig, toProject);
   let copied = 0;
@@ -81,15 +95,22 @@ export async function copyProjectStaging(
     const toLocale = toConfig.locales.find((l) => l.code === locale.code);
     if (!toLocale?.changelog) continue;
 
-    const fromPath = join(ROOT, locale.changelog);
-    const toPath = join(ROOT, toLocale.changelog);
-    try {
-      await readFile(fromPath, "utf-8");
-    } catch {
-      continue;
+    const fromContent = await readStaging(locale.changelog);
+    if (!fromContent.trim()) continue;
+
+    const blocksToMerge: Array<{ label: string; content: string }> = [];
+    for (const version of versionList) {
+      const block = blockForVersion(fromContent, version);
+      if (!block) continue;
+      blocksToMerge.push({ label: block.label, content: block.updateBlock });
     }
+    if (blocksToMerge.length === 0) continue;
+
+    const toPath = join(ROOT, toLocale.changelog);
+    const existing = await readStaging(toLocale.changelog);
+    const output = mergeStagingBlocks(existing, blocksToMerge);
     await mkdir(dirname(toPath), { recursive: true });
-    await copyFile(fromPath, toPath);
+    await writeFile(toPath, output);
     copied++;
   }
 

.github/scripts/cms/prepare-cms-changelog.ts

@@ -103,7 +103,7 @@ async function prepareForProject(
   dryRun: boolean,
   force: boolean,
   explicitVersions: string[]
-): Promise<void> {
+): Promise<string[]> {
   const config = configForProject(baseConfig, projectId);
 
   const docsEnPath = join(ROOT, config.docs_en);
@@ -113,7 +113,7 @@ async function prepareForProject(
   const incremental = incrementalVersionSet();
   if (incremental && incremental.size === 0 && explicitVersions.length === 0) {
     console.log(`[${projectId}] No new <Update> blocks — nothing to prepare.`);
-    return;
+    return [];
   }
 
   const targetEntries = resolveTargetVersions(
@@ -148,9 +148,10 @@ async function prepareForProject(
 
   if (entriesToPrepare.length === 0) {
     console.log(`[${projectId}] No versions to prepare (min=${config.min_version}).`);
-    return;
+    return [];
   }
 
+  const preparedVersions = entriesToPrepare.map((e) => e.version);
   const targetVersions = versionLabelsFromEntries(entriesToPrepare);
   const modeLabel = mode === "en-only" ? "simplify EN" : "translate locales";
   console.log(
@@ -172,7 +173,7 @@ async function prepareForProject(
     );
     console.log(`[${projectId}] Staging: ${enStagingPath}`);
     console.log(`[${projectId}] Next: review staging EN, then pnpm cms:prepare:locales`);
-    return;
+    return preparedVersions;
   }
 
   // translate-only — read existing simplified EN staging (never re-simplify)
@@ -212,6 +213,7 @@ async function prepareForProject(
   );
   console.log(`[${projectId}] Staging: ${enStagingPath.replace("/en/", "/{locale}/")}`);
   console.log(`[${projectId}] Next: review staging locales, then pnpm cms:preview / cms:sync`);
+  return preparedVersions;
 }
 
 async function main(): Promise<void> {
@@ -232,14 +234,32 @@ async function main(): Promise<void> {
 
   const [primary, ...secondary] = projects;
 
-  await prepareForProject(baseConfig, primary!, mode, dryRun, force, explicitVersions);
+  const preparedVersions = await prepareForProject(
+    baseConfig,
+    primary!,
+    mode,
+    dryRun,
+    force,
+    explicitVersions
+  );
 
   for (const projectId of secondary) {
-    if (dryRun) {
-      console.log(`[${projectId}] Would copy staging from ${primary}`);
+    if (preparedVersions.length === 0) {
+      console.log(`[${projectId}] Skip staging merge — no versions prepared`);
+    } else if (dryRun) {
+      console.log(
+        `[${projectId}] Would merge v${preparedVersions.join(", v")} from ${primary} (leave other versions untouched)`
+      );
     } else {
-      const copied = await copyProjectStaging(baseConfig, primary!, projectId);
-      console.log(`[${projectId}] Copied ${copied} staging file(s) from ${primary}`);
+      const copied = await copyProjectStaging(
+        baseConfig,
+        primary!,
+        projectId,
+        preparedVersions
+      );
+      console.log(
+        `[${projectId}] Merged v${preparedVersions.join(", v")} into ${copied} staging file(s) from ${primary} (other versions untouched)`
+      );
     }
     if (projects.length > 1) console.log("");
   }

.github/scripts/cms/published-versions.json

@@ -595,6 +595,34 @@
         "zh"
       ],
       "published_at": "2026-08-08T05:14:50.243Z"
+    },
+    {
+      "project": "cloud",
+      "version": "0.32.0",
+      "locales": [
+        "en",
+        "es",
+        "fr",
+        "ja",
+        "ko",
+        "ru",
+        "zh"
+      ],
+      "published_at": "2026-08-12T15:50:17.011Z"
+    },
+    {
+      "project": "comfyui",
+      "version": "0.32.0",
+      "locales": [
+        "en",
+        "es",
+        "fr",
+        "ja",
+        "ko",
+        "ru",
+        "zh"
+      ],
+      "published_at": "2026-08-12T15:49:27.833Z"
     }
   ]
 }

.github/scripts/cms/staging/cloud/en/changelog/index.mdx

@@ -4,6 +4,18 @@
 ---
 
 
+<Update label="v0.32.0" description="August 11, 2026">
+
+**New Open-Source Model Support**
+* [**LTX 2.5**](https://links.comfy.org/cloLTX2_5): Native support with STG, dual CFG, duration prediction
+
+**Partner Node Updates**
+* [**Qwen Image 3.0**](https://links.comfy.org/45W0b6L): Adds 3.0 and 3.0 Pro text-to-image and edit nodes
+* [**LTX 2.5**](https://links.comfy.org/3TV1HDB): Text, image, and audio to video partner nodes
+* [**Grok Imagine Image 2.0**](https://github.com/Comfy-Org/ComfyUI/pull/15496): Adds support for the grok-imagine-image-2.0 model
+
+</Update>
+
 <Update label="v0.31.0" description="August 7, 2026">
 
 **New Open-Source Model Support**
@@ -11,8 +23,8 @@
 
 **Partner Node Updates**
 * [**ByteDance SeeDance 2.5**](https://links.comfy.org/3RPYzbv): Added ByteDance SeeDance 2.5 model support
 * [**Seedream 5.0 Layer Separation**](https://links.comfy.org/4hgRB9L): ByteDance Seedream 5.0 Pro Layer Separation Support
 * [**Reve deprecated**](https://github.com/Comfy-Org/ComfyUI/pull/15331): Reve partner nodes have been deprecated
 
 **New Node Updates**
 * [**Create Layered Image**](https://github.com/Comfy-Org/ComfyUI/pull/15317): PSD-style layer compositing with blend modes and layer state
@@ -28,7 +40,7 @@
 **Partner Node Updates**
 * [**BFL Flux 3 Video**](https://links.comfy.org/3TVdfGP): Added Flux 3 video model support
 * [**TopazAI Bloom 2 & Wonder 3.5**](https://github.com/Comfy-Org/ComfyUI/pull/15294): Added Bloom 2 and Wonder 3.5 model support
 * [**Kling EOL**](https://github.com/Comfy-Org/ComfyUI/pull/15249): Removed retired legacy Kling models and Virtual Try-On API
 
 </Update>
 
@@ -43,12 +55,12 @@
 
 **New Open-Source Model Support**
 * [**MiniMax-H3**](https://links.comfy.org/3TuT9TO): Adds audio-video model support for MiniMax-H3
 * [**Pruna LTX VAE**](https://github.com/Comfy-Org/ComfyUI/pull/15129): PrunaVAED support for faster LTX 2.3 decoding
 
 **New Node Updates**
 * [**Save Video CRF**](https://github.com/Comfy-Org/ComfyUI/pull/15191): Adds CRF option to Save Video node
 * [**SaveText CSV**](https://github.com/Comfy-Org/ComfyUI/pull/15217): SaveText node can now save .csv output
 * [**VAEDecodeAudio**](https://github.com/Comfy-Org/ComfyUI/pull/15211): Handles decoding of nested audio latents
 
 **Partner Node Updates**
 * [**xAI Grok Imagine Video 1.5**](https://github.com/Comfy-Org/ComfyUI/pull/15197): Updated xAI nodes for grok-imagine-video-1.5 support
@@ -71,13 +83,13 @@
 * [**Gemma4 12B**](https://github.com/Comfy-Org/ComfyUI/pull/14304): Gemma4 12B model support
 * [**MageFlow**](https://github.com/Comfy-Org/ComfyUI/pull/15026): MageFlow model support
 * [**Uni3C ControlNet**](https://github.com/Comfy-Org/ComfyUI/pull/14946): Native Uni3C ControlNet support for Wan models
 * [**Anima**](https://github.com/Comfy-Org/ComfyUI/pull/14954): Anima llite control model support
 
 **New Node Updates**
 * [**Trainer video processing**](https://github.com/Comfy-Org/ComfyUI/pull/13588): Video processing nodes and Image Processing Node video support
 
 **Partner Node Updates**
 * [**Recraft V4.1**](https://github.com/Comfy-Org/ComfyUI/pull/15105): Recraft V4.1 models and V4 image edit nodes
 * [**ByteDance audio multilingual**](https://github.com/Comfy-Org/ComfyUI/pull/15034): ByteDance audio-1.0-multilingual model support
 * [**OpenRouter**](https://github.com/Comfy-Org/ComfyUI/pull/15021): New OpenRouter models, including Claude Opus 5
 * [**Anthropic**](https://github.com/Comfy-Org/ComfyUI/pull/15023): New Anthropic models, including Claude Opus 5
@@ -87,7 +99,7 @@
 <Update label="v0.28.3" description="July 22, 2026">
 
 **Partner Node Updates**
 * [**Gemini Video Omni**](https://github.com/Comfy-Org/ComfyUI/pull/15014): Pass videos as inline data in Gemini Omni partner node
 * [**OpenRouter**](https://github.com/Comfy-Org/ComfyUI/pull/15021): New OpenRouter models
 * [**Anthropic**](https://github.com/Comfy-Org/ComfyUI/pull/15023): New Anthropic models
 
@@ -96,7 +108,7 @@
 <Update label="v0.28.2" description="July 18, 2026">
 
 **Partner Node Updates**
 * [**Gemini Video Omni**](https://github.com/Comfy-Org/ComfyUI/pull/14984): Switched Omni model to the Interactions API
 
 </Update>
 
@@ -112,19 +124,19 @@
 <Update label="v0.28.0" description="July 15, 2026">
 
 **New Open-Source Model Support**
 * [**SeedVR2**](https://github.com/Comfy-Org/workflow_templates/blob/main/templates/utility_seedvr2_3b_int8_upscale_image.json): Native SeedVR2 image and video upscaling
 * [**PID 1.5**](https://github.com/Comfy-Org/ComfyUI/pull/14894): Added PixelDiT PID 1.5 model support
 * [**int4 Models**](https://github.com/Comfy-Org/ComfyUI/pull/14859): convrot int4 support and Turing int8/int4 optimizations
 
 **New Node Updates**
 * [**Save 3D (Advanced)**](https://github.com/Comfy-Org/ComfyUI/pull/14701): Save 3D (Advanced), Save Splat, and Save Point Cloud nodes
 * [**SaveText**](https://github.com/Comfy-Org/ComfyUI/pull/14102): Save text outputs to .txt, .md, or .json
 * [**Text Overlay**](https://github.com/Comfy-Org/ComfyUI/pull/14610): Text overlay for images and video
 * [**Text**](https://github.com/Comfy-Org/ComfyUI/pull/14870): Built-in Text node re-enabled after undeprecation
 
 **Partner Node Updates**
 * [**sync.so sync-3**](https://github.com/Comfy-Org/ComfyUI/pull/14928): sync.so sync-3 lip-sync and talking-image support
 * [**Seedream thinking control**](https://github.com/Comfy-Org/ComfyUI/pull/14853): Option to disable thinking on Seedream nodes
 * [**Gemini Image**](https://github.com/Comfy-Org/ComfyUI/pull/14917): Gemini Image preview models moved to release versions
 
 </Update>
@@ -132,7 +144,7 @@
 <Update label="v0.27.1" description="July 8, 2026">
 
 **Partner Node Updates**
 * [**Seedream 5.0 Pro**](https://links.comfy.org/4ydIJYi): ByteDance image generation model support added.
 * [**Seed Audio 1.0**](https://github.com/Comfy-Org/ComfyUI/pull/14731): Speech, music, sound effects, multi-speaker dialogue.
 * [**Ideogram V1/V2 deprecated**](https://github.com/Comfy-Org/ComfyUI/pull/14712): Legacy Ideogram partner nodes deprecated.
 * [**StabilityAI temporarily removed**](https://github.com/Comfy-Org/ComfyUI/pull/14737): StabilityAI partner nodes temporarily removed.
@@ -142,15 +154,15 @@
 <Update label="v0.27.0" description="June 30, 2026">
 
 **Int8 Support**
 * [**int8 convrot models**](https://github.com/Comfy-Org/ComfyUI/pull/14636): Support for int8 quantized convrot models with faster inference on Turing GPUs
 
 **Partner Node Updates**
 * [**Nano Banana 2 Lite**](https://github.com/Comfy-Org/ComfyUI/pull/14693): Google's lightweight image generation model
 * [**Gemini Video Omni**](https://github.com/Comfy-Org/ComfyUI/pull/14695): Video generation node powered by Gemini
 
 **New Node Updates**
 * [**Seed Node**](https://github.com/Comfy-Org/ComfyUI/pull/14616): Simplifies seed management for reproducible workflows
 * [**Advanced Krea 2 Model Merging**](https://github.com/Comfy-Org/ComfyUI/pull/14621): Merges Krea 2 models
 * [**Bounding Box Canvas**](https://github.com/Comfy-Org/ComfyUI/pull/14537): Canvas with Ideogram JSON prompt support
 * [**ConditioningMultiply**](https://github.com/Comfy-Org/ComfyUI/pull/14686): Multiplies conditioning values
 * [**GLSL Update**](https://github.com/Comfy-Org/ComfyUI/pull/13195): GLSL node updated to use ANGLE library
@@ -168,7 +180,7 @@
 
 **Partner Node Updates**
 * [**SeeDance 2.0**](https://github.com/Comfy-Org/ComfyUI/pull/14614): Added 4K resolution support for ByteDance SeeDance 2.0
 * [**HappyHorse 1.1**](https://links.comfy.org/3SoBFYI): Added support for Alibaba's HappyHorse 1.1 model
 * [**Grok Video**](https://github.com/Comfy-Org/ComfyUI/pull/14612): Added 1080p support for Grok video generation
 
 </Update>
@@ -176,24 +188,24 @@
 <Update label="v0.26.0" description="June 23, 2026">
 
 **New Open-Source Model Support**
 * [**Krea2**](https://links.comfy.org/4eC8714): Krea's open-source image model with wide aesthetic and stylistic range
 * [**Boogu-Image**](https://github.com/Comfy-Org/ComfyUI/pull/14523): Text-to-image with Qwen3-VL-8B encoder and image editing
 * [**Qwen3-VL Text Generation**](https://github.com/Comfy-Org/ComfyUI/pull/14298): Image-aware text generation with 4B and 8B models
 * [**SCAIL-2 Multireference**](https://github.com/Comfy-Org/ComfyUI/pull/14509): Multi-reference SCAIL-2 character replacement
 
 **New Node Updates**
 * [**LTX2 Context Windows**](https://github.com/Comfy-Org/ComfyUI/pull/13325): Context window sampling for LTX-2 video+audio with IC-LoRA splicing
 * [**Save Nodes Output Socket**](https://github.com/Comfy-Org/ComfyUI/pull/13866): Output socket on save nodes for downstream wiring
 
 **Partner Node Updates**
 * [**Luma Rays 3.2**](https://github.com/Comfy-Org/ComfyUI/pull/14540): Seven nodes for Luma Ray 3.2 video generation
 
 </Update>
 
 <Update label="v0.25.1" description="June 16, 2026">
 
 **Partner Node Updates**
 * [**Kling V3-Turbo**](https://github.com/Comfy-Org/ComfyUI/pull/14528): Text-to-video generation model
 
 </Update>
 
@@ -213,9 +225,9 @@
 <Update label="v0.24.1" description="June 3, 2026">
 
 **Partner Node Updates**
 * **Bria Green Background**: Image green background removal node.
 * **Bria Transparent Video Background**: Video background removal with alpha channel output.
 * **Krea 2 Medium Turbo**: Faster medium-sized image generation model.
 * **Seed input to Flux Erase node**: Adds seed parameter for repeatable erasure results.
 
 </Update>
@@ -226,7 +238,7 @@
 * [**Microsoft Lens**](https://github.com/Comfy-Org/ComfyUI/pull/14077): Text-to-image model with GPT-OSS-20B encoder
 * [**NVIDIA PixelDiT & PiD**](https://github.com/Comfy-Org/ComfyUI/pull/14103): Text-to-image models including PixelDiT 1300M and PiD family
 * [**TripoSplat**](https://github.com/Comfy-Org/ComfyUI/pull/14210): Image-to-3D Gaussian splat model
 * [**MediaPipe Face Detection**](https://github.com/Comfy-Org/ComfyUI/pull/14009): Dependency-free face detection and landmarker
 
 Partner Node Updates
 * **OpenRouter LLM Node**: Partner node for OpenRouter language model API
@@ -237,7 +249,7 @@
 
 **Partner Node Updates**
 * **Krea2 Image nodes**: Added new image generation nodes for Krea2 workflows
 * **SeeDance 2**: Improved video reference uploading via memoryview
 
 </Update>
 
@@ -253,7 +265,7 @@
 
 * **OpenRouter LLM node**: New partner node for OpenRouter API access.
 * **Anthropic node**: Added reasoning widget for step-by-step thought display.
 * **ByteDance2Reference node**: Added automatic upscaling widget for video outputs.
 * **Grok LLM**: Fixed issue passing images to the model.
 
 </Update>

.github/scripts/cms/staging/cloud/es/changelog/index.mdx

@@ -4,256 +4,268 @@
 ---
 
 
+<Update label="v0.32.0" description="11 de agosto de 2026">
+
+**Nuevo soporte para modelos de código abierto**
+* [**LTX 2.5**](https://links.comfy.org/cloLTX2_5): Soporte nativo con STG, CFG dual y predicción de duración
+
+**Actualizaciones de nodos asociados**
+* [**Qwen Image 3.0**](https://links.comfy.org/45W0b6L): Añade nodos de texto a imagen y edición para 3.0 y 3.0 Pro
+* [**LTX 2.5**](https://links.comfy.org/3TV1HDB): Nodos asociados de texto, imagen y audio a video
+* [**Grok Imagine Image 2.0**](https://github.com/Comfy-Org/ComfyUI/pull/15496): Añade soporte para el modelo grok-imagine-image-2.0
+
+</Update>
+
 <Update label="v0.31.0" description="7 de agosto de 2026">
 
 **Nuevo soporte para modelos de código abierto**
 * [**Wan-Animate2**](https://links.comfy.org/3TAWMHV): Wan-Animate2 nativo con controles de pose/referencia y caché opcional de rama de pose
 
 **Actualizaciones de nodos partner**
 * [**ByteDance SeeDance 2.5**](https://links.comfy.org/3RPYzbv): Se añadió soporte para el modelo ByteDance SeeDance 2.5
 * [**Seedream 5.0 Layer Separation**](https://links.comfy.org/4hgRB9L): Soporte de separación de capas de ByteDance Seedream 5.0 Pro
 * [**Reve obsoleto**](https://github.com/Comfy-Org/ComfyUI/pull/15331): Los nodos partner de Reve han quedado obsoletos
 
 **Actualizaciones de nuevos nodos**
 * [**Create Layered Image**](https://github.com/Comfy-Org/ComfyUI/pull/15317): Composición de capas estilo PSD con modos de fusión y estado de capa
 * [**Add Layer**](https://github.com/Comfy-Org/ComfyUI/pull/15317): Añadir una capa a una pila de capas del compositor
 * [**Layers From Bounding Boxes**](https://github.com/Comfy-Org/ComfyUI/pull/15317): Construir capas del compositor a partir de entradas de cuadros delimitadores
 * [**WanAnimate2ToVideo**](https://github.com/Comfy-Org/ComfyUI/pull/15362): Acondicionamiento de Wan-Animate2 con controles de pose y referencia
 * [**WanAnimate2Cache**](https://github.com/Comfy-Org/ComfyUI/pull/15362): Caché opcional de rama de pose para acelerar Wan-Animate2
 
 </Update>
 
 <Update label="v0.30.2" description="5 de agosto de 2026">
 
 **Actualizaciones de nodos de socios**
 * [**BFL Flux 3 Video**](https://links.comfy.org/3TVdfGP): Añadido soporte para el modelo de video Flux 3
 * [**TopazAI Bloom 2 y Wonder 3.5**](https://github.com/Comfy-Org/ComfyUI/pull/15294): Añadido soporte para los modelos Bloom 2 y Wonder 3.5
 * [**Kling EOL**](https://github.com/Comfy-Org/ComfyUI/pull/15249): Eliminados los modelos Kling heredados retirados y la API de Virtual Try-On
 
 </Update>
 
 <Update label="v0.30.1" description="3 de agosto de 2026">
 
 **Rendimiento y estabilidad**
 * [**Frontend**](https://github.com/Comfy-Org/ComfyUI/pull/15244): Se actualizó comfyui-frontend-package a 1.47.12
 
 </Update>
 
 <Update label="v0.30.0" description="3 de agosto de 2026">
 
 **Nuevo soporte de modelos de código abierto**
 * [**MiniMax-H3**](https://links.comfy.org/3TuT9TO): Añade soporte para el modelo de audio y vídeo MiniMax-H3
 * [**Pruna LTX VAE**](https://github.com/Comfy-Org/ComfyUI/pull/15129): Soporte de PrunaVAED para una decodificación LTX 2.3 más rápida
 
 **Nuevas actualizaciones de nodos**
 * [**Save Video CRF**](https://github.com/Comfy-Org/ComfyUI/pull/15191): Añade la opción CRF al nodo Save Video
 * [**SaveText CSV**](https://github.com/Comfy-Org/ComfyUI/pull/15217): El nodo SaveText ahora puede guardar la salida en .csv
 * [**VAEDecodeAudio**](https://github.com/Comfy-Org/ComfyUI/pull/15211): Gestiona la decodificación de latentes de audio anidados
 
 **Actualizaciones de nodos de socios**
 * [**xAI Grok Imagine Video 1.5**](https://github.com/Comfy-Org/ComfyUI/pull/15197): Nodos xAI actualizados para dar soporte a grok-imagine-video-1.5
 * [**MiniMax H3 768P**](https://github.com/Comfy-Org/ComfyUI/pull/15227): Añade la resolución 768P para los nodos MiniMax H3
 
 </Update>
 
 <Update label="v0.29.2" description="31 de julio de 2026">
 
 **Actualizaciones de nodos de socios**
 * [**Ideogram P-Image**](https://github.com/Comfy-Org/ComfyUI/pull/15154): Añade soporte para el modelo Ideogram P-Image
 * [**MiniMax H3**](https://github.com/Comfy-Org/ComfyUI/pull/15167): Añade soporte para el modelo MiniMax H3
 
 </Update>
 
 <Update label="v0.29.0" description="29 de julio de 2026">
 
 **Nuevo soporte para modelos de código abierto**
 * [**JoyImageEdit**](https://github.com/Comfy-Org/ComfyUI/pull/14428): Soporte nativo para el modelo JoyImageEdit
 * [**Gemma4 12B**](https://github.com/Comfy-Org/ComfyUI/pull/14304): Soporte para el modelo Gemma4 12B
 * [**MageFlow**](https://github.com/Comfy-Org/ComfyUI/pull/15026): Soporte para el modelo MageFlow
 * [**Uni3C ControlNet**](https://github.com/Comfy-Org/ComfyUI/pull/14946): Soporte nativo para Uni3C ControlNet en modelos Wan
 * [**Anima**](https://github.com/Comfy-Org/ComfyUI/pull/14954): Soporte para el modelo de control Anima llite
 
 **Nuevas actualizaciones de nodos**
 * [**Procesamiento de video para Trainer**](https://github.com/Comfy-Org/ComfyUI/pull/13588): Nodos de procesamiento de video y soporte de video en el nodo de procesamiento de imágenes
 
 **Actualizaciones de nodos de socios**
 * [**Recraft V4.1**](https://github.com/Comfy-Org/ComfyUI/pull/15105): Modelos Recraft V4.1 y nodos de edición de imágenes V4
 * [**ByteDance audio multilingüe**](https://github.com/Comfy-Org/ComfyUI/pull/15034): Soporte para el modelo ByteDance audio-1.0-multilingual
 * [**OpenRouter**](https://github.com/Comfy-Org/ComfyUI/pull/15021): Nuevos modelos de OpenRouter, incluido Claude Opus 5
 * [**Anthropic**](https://github.com/Comfy-Org/ComfyUI/pull/15023): Nuevos modelos de Anthropic, incluido Claude Opus 5
 
 </Update>
 
 <Update label="v0.28.3" description="22 de julio de 2026">
 
 **Actualizaciones de nodos partner**
 * [**Gemini Video Omni**](https://github.com/Comfy-Org/ComfyUI/pull/15014): Pasa videos como datos en línea en el nodo partner de Gemini Omni
 * [**OpenRouter**](https://github.com/Comfy-Org/ComfyUI/pull/15021): Nuevos modelos de OpenRouter
 * [**Anthropic**](https://github.com/Comfy-Org/ComfyUI/pull/15023): Nuevos modelos de Anthropic
 
 </Update>
 
 <Update label="v0.28.2" description="July 18, 2026">
 
 **Actualizaciones de nodos asociados**
 * [**Gemini Video Omni**](https://github.com/Comfy-Org/ComfyUI/pull/14984): Cambió el modelo Omni a la API de Interactions
 
 </Update>
 
 <Update label="v0.28.1" description="17 de julio de 2026">
 
 **Actualizaciones de nodos asociados**
 * [**GPT-5.6**](https://github.com/Comfy-Org/ComfyUI/pull/14957): Se añadieron los modelos GPT-5.6 Sol, Terra y Luna
 * [**Gemini 3.5 Flash**](https://github.com/Comfy-Org/ComfyUI/pull/14972): Se añadió Gemini 3.5 Flash al nodo Gemini LLM
 * [**HeyGen**](https://github.com/Comfy-Org/ComfyUI/pull/14958): Nodos de Video de Avatar, Foto Parlante, Crear Avatar, Traducir Video y Texto a Voz
 
 </Update>
 
 <Update label="v0.28.0" description="July 15, 2026">
 
 **Nuevo soporte para modelos de código abierto**
 * [**SeedVR2**](https://github.com/Comfy-Org/workflow_templates/blob/main/templates/utility_seedvr2_3b_int8_upscale_image.json): Escalado nativo de imágenes y vídeos con SeedVR2
 * [**PID 1.5**](https://github.com/Comfy-Org/ComfyUI/pull/14894): Se añadió soporte para el modelo PixelDiT PID 1.5
 * [**int4 Models**](https://github.com/Comfy-Org/ComfyUI/pull/14859): Soporte para convrot int4 y optimizaciones Turing int8/int4
 
 **Nuevas actualizaciones de nodos**
 * [**Save 3D (Advanced)**](https://github.com/Comfy-Org/ComfyUI/pull/14701): Nodos Guardar 3D (Avanzado), Guardar Splat y Guardar Nube de Puntos
 * [**SaveText**](https://github.com/Comfy-Org/ComfyUI/pull/14102): Guardar salidas de texto en .txt, .md o .json
 * [**Text Overlay**](https://github.com/Comfy-Org/ComfyUI/pull/14610): Superposición de texto para imágenes y vídeo
 * [**Text**](https://github.com/Comfy-Org/ComfyUI/pull/14870): Nodo Text integrado rehabilitado
 
 **Actualizaciones de nodos asociados**
 * [**sync.so sync-3**](https://github.com/Comfy-Org/ComfyUI/pull/14928): Soporte para sincronización de labios y talking-image con sync.so sync-3
 * [**Seedream thinking control**](https://github.com/Comfy-Org/ComfyUI/pull/14853): Opción para desactivar el pensamiento en nodos Seedream
 * [**Gemini Image**](https://github.com/Comfy-Org/ComfyUI/pull/14917): Modelos de previsualización de Gemini Image trasladados a versiones de lanzamiento
 
 </Update>
 
 <Update label="v0.27.1" description="July 8, 2026">
 
 **Actualizaciones de nodos asociados**
 * [**Seedream 5.0 Pro**](https://links.comfy.org/4ydIJYi): Soporte para modelo de generación de imágenes de ByteDance añadido.
 * [**Seed Audio 1.0**](https://github.com/Comfy-Org/ComfyUI/pull/14731): Voz, música, efectos de sonido, diálogo multi-orador.
 * [**Ideogram V1/V2 obsoleto**](https://github.com/Comfy-Org/ComfyUI/pull/14712): Nodos asociados heredados de Ideogram obsoletos.
 * [**StabilityAI eliminado temporalmente**](https://github.com/Comfy-Org/ComfyUI/pull/14737): Nodos asociados de StabilityAI eliminados temporalmente.
 
 </Update>
 
 <Update label="v0.27.0" description="30 de junio de 2026">
 
 **Soporte Int8**
 * [**modelos convrot int8**](https://github.com/Comfy-Org/ComfyUI/pull/14636): Soporte para modelos convrot cuantificados int8 con inferencia más rápida en GPUs Turing
 
 **Actualizaciones de Nodos Asociados**
 * [**Nano Banana 2 Lite**](https://github.com/Comfy-Org/ComfyUI/pull/14693): El modelo ligero de generación de imágenes de Google
 * [**Gemini Video Omni**](https://github.com/Comfy-Org/ComfyUI/pull/14695): Nodo de generación de video impulsado por Gemini
 
 **Actualizaciones de Nuevos Nodos**
 * [**Seed Node**](https://github.com/Comfy-Org/ComfyUI/pull/14616): Simplifica la gestión de semillas para flujos de trabajo reproducibles
 * [**Advanced Krea 2 Model Merging**](https://github.com/Comfy-Org/ComfyUI/pull/14621): Fusiona modelos Krea 2
 * [**Bounding Box Canvas**](https://github.com/Comfy-Org/ComfyUI/pull/14537): Lienzo con soporte de prompt JSON de Ideogram
 * [**ConditioningMultiply**](https://github.com/Comfy-Org/ComfyUI/pull/14686): Multiplica valores de condicionamiento
 * [**GLSL Update**](https://github.com/Comfy-Org/ComfyUI/pull/13195): Nodo GLSL actualizado para usar la biblioteca ANGLE
 
 </Update>
 
 <Update label="v0.26.2" description="25 de junio de 2026">
 
 **Actualizaciones de nodos Partner**
 * [**SeeDance 2.0 Mini**](https://links.comfy.org/4eOUnjJ): Se añadió ByteDance SeeDance 2.0 Mini en los nodos Text-to-Video, First-Last-Frame y Reference
 
 </Update>
 
 <Update label="v0.26.1" description="24 de junio de 2026">
 
 **Actualizaciones de nodos asociados**
 * [**SeeDance 2.0**](https://github.com/Comfy-Org/ComfyUI/pull/14614): Se añadió compatibilidad con resolución 4K para ByteDance SeeDance 2.0
 * [**HappyHorse 1.1**](https://links.comfy.org/3SoBFYI): Se añadió compatibilidad con el modelo HappyHorse 1.1 de Alibaba
 * [**Grok Video**](https://github.com/Comfy-Org/ComfyUI/pull/14612): Se añadió compatibilidad con 1080p para la generación de video Grok
 
 </Update>
 
 <Update label="v0.26.0" description="23 de junio de 2026">
 
 **Nuevo soporte para modelos de código abierto**
 * [**Krea2**](https://links.comfy.org/4eC8714): Modelo de imagen de código abierto de Krea con amplio rango estético y estilístico
 * [**Boogu-Image**](https://github.com/Comfy-Org/ComfyUI/pull/14523): Texto a imagen con codificador Qwen3-VL-8B y edición de imágenes
 * [**Generación de texto Qwen3-VL**](https://github.com/Comfy-Org/ComfyUI/pull/14298): Generación de texto consciente de la imagen con modelos 4B y 8B
 * [**SCAIL-2 Multireferencia**](https://github.com/Comfy-Org/ComfyUI/pull/14509): Reemplazo de personajes con múltiples referencias SCAIL-2
 
 **Nuevas actualizaciones de nodos**
 * [**Ventanas de contexto LTX2**](https://github.com/Comfy-Org/ComfyUI/pull/13325): Muestreo de ventana de contexto para video+audio LTX-2 con empalme IC-LoRA
 * [**Socket de salida de nodos de guardado**](https://github.com/Comfy-Org/ComfyUI/pull/13866): Socket de salida en nodos de guardado para cableado descendente
 
 **Actualizaciones de nodos de socios**
 * [**Luma Rays 3.2**](https://github.com/Comfy-Org/ComfyUI/pull/14540): Siete nodos para la generación de video Luma Ray 3.2
 
 </Update>
 
 <Update label="v0.25.1" description="16 de junio de 2026">
 
 **Actualizaciones de Nodos Asociados**
 * [**Kling V3-Turbo**](https://github.com/Comfy-Org/ComfyUI/pull/14528): Modelo de generación de texto a video
 
 </Update>
 
 <Update label="v0.25.0" description="June 16, 2026">
 
 **Soporte para nuevos modelos de código abierto**
 * [**Depth Anything 3**](https://github.com/Comfy-Org/ComfyUI/pull/13853): Modelo de estimación de profundidad monocular
 * [**SCAIL-2**](https://github.com/Comfy-Org/ComfyUI/pull/14373): Modelo de reemplazo de personajes para edición de video
 * [**Bernini-R**](https://github.com/Comfy-Org/ComfyUI/pull/14216): Soporte para modelo de video Wan
 
 **Actualizaciones de nodos de socios**
 * [**Gemini Text Node**](https://github.com/Comfy-Org/ComfyUI/pull/14299): Nodo de socio LLM
 * [**Runway Aleph2**](https://github.com/Comfy-Org/ComfyUI/pull/14306): Nodo de generación de video
 
 </Update>
 
 <Update label="v0.24.1" description="3 de junio de 2026">
 
 **Actualizaciones de nodos asociados**
 * **Bria Green Background**: Nodo de eliminación de fondo verde de imagen.
 * **Bria Transparent Video Background**: Eliminación de fondo de video con salida de canal alfa.
 * **Krea 2 Medium Turbo**: Modelo más rápido de generación de imágenes de tamaño mediano.
 * **Entrada de semilla al nodo Flux Erase**: Añade parámetro de semilla para resultados de borrado repetibles.
 
 </Update>
 
 <Update label="v0.23.0" description="1 de junio de 2026">
 
 Nuevo soporte para modelos de código abierto
 * [**Microsoft Lens**](https://github.com/Comfy-Org/ComfyUI/pull/14077): Modelo de texto a imagen con codificador GPT-OSS-20B
 * [**NVIDIA PixelDiT & PiD**](https://github.com/Comfy-Org/ComfyUI/pull/14103): Modelos de texto a imagen que incluyen PixelDiT 1300M y la familia PiD
 * [**TripoSplat**](https://github.com/Comfy-Org/ComfyUI/pull/14210): Modelo de splats gaussianos de imagen a 3D
 * [**MediaPipe Face Detection**](https://github.com/Comfy-Org/ComfyUI/pull/14009): Detección de rostros y puntos de referencia sin dependencias
 
 Actualizaciones de nodos asociados
 * **OpenRouter LLM Node**: Nodo asociado para la API de modelo de lenguaje de OpenRouter
 
 </Update>
 
 <Update label="v0.22.3" description="May 20, 2026">
 
 **Actualizaciones de nodos asociados**
 * **Nodos de imagen Krea2**: Se añadieron nuevos nodos de generación de imágenes para flujos de trabajo de Krea2
 * **SeeDance 2**: Se mejoró la carga de referencias de video mediante memoryview
 
 </Update>
 
 <Update label="v0.22.2" description="20 de mayo de 2026">
 
 **Actualizaciones de nodos asociados**
 * **Rodin2.5**: Nodo de generación de malla de calidad
 * **Rodin2.5**: Corrección en el manejo del parámetro semilla
 
 </Update>
 
 <Update label="v0.22.1" description="20 de mayo de 2026">
 
 * **OpenRouter LLM node**: Nuevo nodo asociado para el acceso a la API de OpenRouter.
 * **Anthropic node**: Añadido widget de razonamiento para mostrar el pensamiento paso a paso.
 * **ByteDance2Reference node**: Añadido widget de escalado automático para las salidas de video.
 * **Grok LLM**: Corregido un problema al pasar imágenes al modelo.
 
 </Update>