From 54bdea7c1b20e8af58106b454c0ed14725478ead Mon Sep 17 00:00:00 2001 From: lin-bot23 Date: Mon, 16 Nov 2026 13:07:51 +0800 Subject: [PATCH] docs(minimax): add a MiniMax H3 Image API page --- docs.json | 30 +++++++ .../minimax/minimax-h3-image.mdx | 79 +++++++++++++++++++ .../minimax/minimax-h3-image.mdx | 79 +++++++++++++++++++ .../minimax/minimax-h3-image.mdx | 79 +++++++++++++++++++ .../minimax/minimax-h3-image.mdx | 71 +++++++++++++++++ .../minimax/minimax-h3-image.mdx | 79 +++++++++++++++++++ 6 files changed, 417 insertions(+) create mode 100644 es/tutorials/partner-nodes/minimax/minimax-h3-image.mdx create mode 100644 ja/tutorials/partner-nodes/minimax/minimax-h3-image.mdx create mode 100644 ko/tutorials/partner-nodes/minimax/minimax-h3-image.mdx create mode 100644 tutorials/partner-nodes/minimax/minimax-h3-image.mdx create mode 100644 zh/tutorials/partner-nodes/minimax/minimax-h3-image.mdx diff --git a/docs.json b/docs.json index 057269edf9..3b80746d40 100644 --- a/docs.json +++ b/docs.json @@ -675,6 +675,12 @@ "tutorials/partner-nodes/meta/muse-image" ] }, + { + "group": "MiniMax", + "pages": [ + "tutorials/partner-nodes/minimax/minimax-h3-image" + ] + }, { "group": "OpenAI", "pages": [ @@ -4950,6 +4956,12 @@ "zh/tutorials/partner-nodes/meta/muse-image" ] }, + { + "group": "MiniMax", + "pages": [ + "zh/tutorials/partner-nodes/minimax/minimax-h3-image" + ] + }, { "group": "OpenAI", "pages": [ @@ -9158,6 +9170,12 @@ "ja/tutorials/partner-nodes/meta/muse-image" ] }, + { + "group": "MiniMax", + "pages": [ + "ja/tutorials/partner-nodes/minimax/minimax-h3-image" + ] + }, { "group": "OpenAI", "pages": [ @@ -13383,6 +13401,12 @@ "ko/tutorials/partner-nodes/meta/muse-image" ] }, + { + "group": "MiniMax", + "pages": [ + "ko/tutorials/partner-nodes/minimax/minimax-h3-image" + ] + }, { "group": "OpenAI", "pages": [ @@ -17528,6 +17552,12 @@ "es/tutorials/partner-nodes/meta/muse-image" ] }, + { + "group": "MiniMax", + "pages": [ + "es/tutorials/partner-nodes/minimax/minimax-h3-image" + ] + }, { "group": "OpenAI", "pages": [ diff --git a/es/tutorials/partner-nodes/minimax/minimax-h3-image.mdx b/es/tutorials/partner-nodes/minimax/minimax-h3-image.mdx new file mode 100644 index 0000000000..2e04201565 --- /dev/null +++ b/es/tutorials/partner-nodes/minimax/minimax-h3-image.mdx @@ -0,0 +1,79 @@ +--- +title: "MiniMax H3 Image API: Texto a imagen y edición de imágenes en ComfyUI" +description: "Genera imágenes 2K o 4K desde un prompt o edita hasta 16 imágenes de referencia con MiniMax H3 Image mediante la API de socios de ComfyUI." +sidebarTitle: "MiniMax H3 Image" +translationSourceHash: ffae3038 +translationFrom: tutorials/partner-nodes/minimax/minimax-h3-image.mdx +translationBlockHashes: + "_intro": b2e9cda8 + "What MiniMax H3 Image is good at": 028a76d3 + "Text to image {#api_minimax_h3_image_t2i}": cd37d4b1 + "Image edit {#api_minimax_h3_image_image_edit}": cacb7043 + "Options": b178c63a + "Get started": d289ab9b +--- +import ReqHint from "/snippets/es/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/es/tutorials/update-reminder.mdx"; + +MiniMax H3 Image es el modelo de imagen de MiniMax, disponible en ComfyUI a través del nodo de socio **MiniMax H3 Image**. Texto a imagen y la edición de imágenes usan el mismo nodo: sin imágenes de referencia conectadas, genera a partir de tu prompt, y con referencias conectadas, las edita o las combina. + +Los flujos de trabajo se ejecutan en los servidores de MiniMax a través de tu cuenta de Comfy API, así que no hay descargas de modelos ni una GPU local que administrar. + + + + +## En qué destaca MiniMax H3 Image + +- **Salida en 2K y 4K**: elige un nivel de tamaño y una relación de aspecto, o define un ancho y un alto exactos +- **Edición de imágenes**: edita o combina hasta 16 imágenes de referencia, a las que haces referencia en el prompt como `@Image1` a `@Image16` +- **Reescritura de prompts**: de forma predeterminada, MiniMax amplía un prompt corto hasta convertirlo en una descripción detallada de la escena y devuelve el prompt reescrito junto con la imagen +- **Detalle de nivel de cartel**: adecuado para carteles, embalajes e imágenes editoriales que necesitan texto legible y detalle denso + +## Texto a imagen {#api_minimax_h3_image_t2i} + +Genera una imagen a partir de un prompt en 2K o 4K. + +Vista previa del flujo de trabajo de texto a imagen de MiniMax H3 Image + + + + Descarga el JSON o busca "MiniMax-H3-Image: Text to Image (API)" en la Biblioteca de plantillas + + + +## Editar imagen {#api_minimax_h3_image_image_edit} + +Conecta una imagen de origen, describe el cambio y solo se reescribe lo que pide el prompt. La plantilla agrega una mujer a un render finalizado y conserva el resto del encuadre. + +Vista previa del flujo de trabajo de edición de imágenes de MiniMax H3 Image + + + + Descarga el JSON o busca "MiniMax-H3-Image: Image Edit (API)" en la Biblioteca de plantillas + + + +**Materiales de entrada** + + + + `LoadImage` · `surreal_rendering_modern.png` + + + +## Opciones + +| Entrada | Valores | Notas | +| :---- | :----- | :---- | +| `model` | `MiniMax-H3-Image`, `MiniMax-H3-Image-Highspeed` | En la plantilla, cambia el menú desplegable `model` para cambiar de modelo | +| `size` | `2K`, `4K`, `custom` | 2K y 4K derivan el ancho y el alto de la relación de aspecto. `custom` toma un ancho y un alto de 512 a 7680 px cada uno, con 655,360 a 20,000,000 px en total y una relación de aspecto de 3:1 como máximo | +| `aspect_ratio` | `auto` más 15 relaciones fijas | `auto` sigue el prompt, o la imagen de referencia cuando hay una conectada | +| `enhance_prompt` | activado de forma predeterminada | MiniMax reescribe primero el prompt como una descripción detallada de la escena. Las ediciones siempre se mejoran, así que desactivarlo solo afecta a Texto a imagen, que entonces necesita una relación fija o un tamaño personalizado | +| `images` | hasta 16 | Imágenes de referencia para editar y combinar. Menciónalas en el prompt como `@Image1`, `@Image2`, etc., en el orden de conexión | + +Los prompts pueden tener hasta 8,000 caracteres cuando hay imágenes de referencia conectadas. El nodo también devuelve el `revised_prompt` reescrito junto a la imagen. + +## Empezar + +1. Actualiza ComfyUI a 0.45.1 o posterior +2. En la biblioteca de plantillas, busca `MiniMax-H3-Image` diff --git a/ja/tutorials/partner-nodes/minimax/minimax-h3-image.mdx b/ja/tutorials/partner-nodes/minimax/minimax-h3-image.mdx new file mode 100644 index 0000000000..cfae136ff3 --- /dev/null +++ b/ja/tutorials/partner-nodes/minimax/minimax-h3-image.mdx @@ -0,0 +1,79 @@ +--- +title: "MiniMax H3 Image API: ComfyUIでのテキストから画像生成と画像編集" +description: "プロンプトから2Kまたは4K画像を生成するか、ComfyUIパートナーAPIを通じてMiniMax H3 Imageで最大16枚の参照画像を編集します。" +sidebarTitle: "MiniMax H3 Image" +translationSourceHash: ffae3038 +translationFrom: tutorials/partner-nodes/minimax/minimax-h3-image.mdx +translationBlockHashes: + "_intro": b2e9cda8 + "What MiniMax H3 Image is good at": 028a76d3 + "Text to image {#api_minimax_h3_image_t2i}": cd37d4b1 + "Image edit {#api_minimax_h3_image_image_edit}": cacb7043 + "Options": b178c63a + "Get started": d289ab9b +--- +import ReqHint from "/snippets/ja/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/ja/tutorials/update-reminder.mdx"; + +MiniMax H3 Image は MiniMax の画像モデルで、ComfyUI では **MiniMax H3 Image** パートナーノードから利用できます。テキストから画像への生成と画像編集は同じノードを使います。参照画像を接続しない場合はプロンプトから生成し、参照画像を接続した場合はそれらを編集または合成します。 + +ワークフローは Comfy API アカウントを通じて MiniMax のサーバー上で実行されるため、モデルのダウンロードやローカル GPU の管理は不要です。 + + + + +## MiniMax H3 Image が得意なこと + +- **2K・4K 出力**: サイズティアとアスペクト比を選ぶか、幅と高さを正確に指定できます +- **画像編集**: 最大 16 枚の参照画像を編集または結合できます。プロンプト内では `@Image1` から `@Image16` として指定します +- **プロンプトの書き換え**: MiniMax はデフォルトで短いプロンプトを詳細なシーン説明へと展開し、書き換え後のプロンプトを画像とともに返します +- **ポスター品質のディテール**: 読みやすいテキストと緻密なディテールが求められるポスター、パッケージ、エディトリアル画像に適しています + +## テキストから画像を生成 {#api_minimax_h3_image_t2i} + +プロンプトから 2K または 4K の画像を生成します。 + +MiniMax H3 Image テキストから画像へのワークフローのプレビュー + + + + JSON をダウンロードするか、テンプレートライブラリで "MiniMax-H3-Image: Text to Image (API)" を検索してください + + + +## 画像編集 {#api_minimax_h3_image_image_edit} + +ソース画像を接続し、変更内容を記述すると、プロンプトで要求した部分だけが書き換えられます。テンプレートは完成したレンダリングに女性を追加し、フレームの残りの部分は維持します。 + +MiniMax H3 画像編集ワークフローのプレビュー + + + + JSON をダウンロードするか、テンプレートライブラリで「MiniMax-H3-Image: Image Edit (API)」を検索してください + + + +**入力素材** + + + + `LoadImage` · `surreal_rendering_modern.png` + + + +## オプション + +| 入力 | 値 | 備考 | +| :---- | :----- | :---- | +| `model` | `MiniMax-H3-Image`、`MiniMax-H3-Image-Highspeed` | テンプレートでは `model` ドロップダウンを変更して切り替えます | +| `size` | `2K`、`4K`、`custom` | 2K と 4K はアスペクト比から幅と高さを導出します。`custom` は幅と高さをそれぞれ 512〜7680 px の範囲で受け取り、合計は 655,360〜20,000,000 px、アスペクト比は最大 3:1 です | +| `aspect_ratio` | `auto` と 15 種類の固定比率 | `auto` はプロンプトに従うか、参照画像が接続されている場合はその参照画像に従います | +| `enhance_prompt` | デフォルトでオン | MiniMax はまずプロンプトを詳細なシーン記述に書き換えます。編集では常に拡張が行われるため、オフにしても影響するのはテキストからの画像生成のみで、その場合は固定比率またはカスタムサイズが必要です | +| `images` | 最大 16 枚 | 編集や合成のための参照画像です。プロンプト内では接続順に `@Image1`、`@Image2` などのように参照します | + +参照画像が接続されている場合、プロンプトは最大 8,000 文字まで使用できます。ノードは画像とともに、書き換えられた `revised_prompt` も返します。 + +## はじめる + +1. ComfyUIを0.45.1以降にアップデートします +2. テンプレートライブラリで`MiniMax-H3-Image`を検索します diff --git a/ko/tutorials/partner-nodes/minimax/minimax-h3-image.mdx b/ko/tutorials/partner-nodes/minimax/minimax-h3-image.mdx new file mode 100644 index 0000000000..b5f7738862 --- /dev/null +++ b/ko/tutorials/partner-nodes/minimax/minimax-h3-image.mdx @@ -0,0 +1,79 @@ +--- +title: "MiniMax H3 Image API: ComfyUI에서 텍스트 기반 이미지 생성 및 이미지 편집" +description: "프롬프트로 2K 또는 4K 이미지를 생성하거나, ComfyUI 파트너 API를 통해 MiniMax H3 Image로 최대 16개의 참조 이미지를 편집하세요." +sidebarTitle: "MiniMax H3 Image" +translationSourceHash: ffae3038 +translationFrom: tutorials/partner-nodes/minimax/minimax-h3-image.mdx +translationBlockHashes: + "_intro": b2e9cda8 + "What MiniMax H3 Image is good at": 028a76d3 + "Text to image {#api_minimax_h3_image_t2i}": cd37d4b1 + "Image edit {#api_minimax_h3_image_image_edit}": cacb7043 + "Options": b178c63a + "Get started": d289ab9b +--- +import ReqHint from "/snippets/ko/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/ko/tutorials/update-reminder.mdx"; + +MiniMax H3 Image는 MiniMax의 이미지 모델로, ComfyUI에서 **MiniMax H3 Image** 파트너 노드를 통해 사용할 수 있습니다. 텍스트 기반 이미지 생성과 이미지 편집은 동일한 노드를 사용합니다. 참조 이미지가 연결되지 않은 경우 프롬프트로부터 생성하고, 참조 이미지가 연결된 경우에는 이를 편집하거나 결합합니다. + +워크플로는 Comfy API 계정을 통해 MiniMax의 서버에서 실행되므로 모델 다운로드가 없고 관리할 로컬 GPU도 없습니다. + + + + +## MiniMax H3 Image가 잘하는 것 + +- **2K 및 4K 출력**: 크기 등급과 화면 비율을 선택하거나 정확한 너비와 높이를 설정하세요 +- **이미지 편집**: 최대 16개의 참조 이미지를 편집하거나 결합할 수 있으며, 프롬프트에서 `@Image1`부터 `@Image16`까지로 지정합니다 +- **프롬프트 재작성**: MiniMax는 기본적으로 짧은 프롬프트를 상세한 장면 설명으로 확장하고, 재작성된 프롬프트를 이미지와 함께 반환합니다 +- **포스터 수준의 디테일**: 읽기 쉬운 텍스트와 촘촘한 디테일이 필요한 포스터, 패키징, 에디토리얼 이미지에 적합합니다 + +## 텍스트 기반 이미지 생성 {#api_minimax_h3_image_t2i} + +프롬프트로 2K 또는 4K 해상도의 이미지를 생성합니다. + +MiniMax H3 Image 텍스트 기반 이미지 생성 워크플로 미리보기 + + + + JSON을 다운로드하거나 템플릿 라이브러리에서 "MiniMax-H3-Image: Text to Image (API)"를 검색하세요 + + + +## 이미지 편집 {#api_minimax_h3_image_image_edit} + +소스 이미지를 연결하고 변경 사항을 설명하면, 프롬프트가 요청한 부분만 다시 작성됩니다. 템플릿은 완료된 렌더에 여성을 추가하고 나머지 프레임은 그대로 유지합니다. + +MiniMax H3 이미지 편집 워크플로 미리보기 + + + + JSON을 다운로드하거나 템플릿 라이브러리에서 "MiniMax-H3-Image: Image Edit (API)"를 검색하세요 + + + +**입력 자료** + + + + `LoadImage` · `surreal_rendering_modern.png` + + + +## 옵션 + +| 입력 | 값 | 참고 | +| :---- | :----- | :---- | +| `model` | `MiniMax-H3-Image`, `MiniMax-H3-Image-Highspeed` | 템플릿에서 `model` 드롭다운을 변경하여 전환합니다 | +| `size` | `2K`, `4K`, `custom` | 2K와 4K는 가로세로 비율로부터 너비와 높이를 도출합니다. `custom`은 너비와 높이를 각각 512~7680 px로 지정하며, 총 655,360~20,000,000 px이고 가로세로 비율은 최대 3:1입니다 | +| `aspect_ratio` | `auto` 및 고정 비율 15개 | `auto`는 프롬프트를 따르며, 참조 이미지가 연결된 경우에는 해당 참조 이미지를 따릅니다 | +| `enhance_prompt` | 기본적으로 켜짐 | MiniMax는 프롬프트를 먼저 상세한 장면 설명으로 다시 작성합니다. 편집 작업에는 항상 프롬프트 향상이 적용되므로, 이 옵션을 끄면 텍스트 기반 이미지 생성에만 영향을 줍니다. 이 경우 고정 비율이나 사용자 지정 크기가 필요합니다 | +| `images` | 최대 16개 | 편집 및 결합을 위한 참조 이미지입니다. 프롬프트에서 연결된 순서대로 `@Image1`, `@Image2` 등으로 참조하세요 | + +참조 이미지가 연결된 경우 프롬프트는 최대 8,000자까지 사용할 수 있습니다. 노드는 이미지와 함께 다시 작성된 `revised_prompt`도 반환합니다. + +## 시작하기 + +1. ComfyUI를 0.45.1 이상으로 업데이트하세요 +2. Template Library에서 `MiniMax-H3-Image`를 검색하세요 diff --git a/tutorials/partner-nodes/minimax/minimax-h3-image.mdx b/tutorials/partner-nodes/minimax/minimax-h3-image.mdx new file mode 100644 index 0000000000..43a7a6d863 --- /dev/null +++ b/tutorials/partner-nodes/minimax/minimax-h3-image.mdx @@ -0,0 +1,71 @@ +--- +title: "MiniMax H3 Image API: Text to Image and Image Edit in ComfyUI" +description: "Generate 2K or 4K images from a prompt, or edit up to 16 reference images with MiniMax H3 Image through the ComfyUI partner API." +sidebarTitle: "MiniMax H3 Image" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +MiniMax H3 Image is MiniMax's image model, available in ComfyUI through the **MiniMax H3 Image** partner node. Text to image and image editing use the same node: with no reference images connected it generates from your prompt, and with references connected it edits or combines them. + +The workflows run on MiniMax's servers through your Comfy API account, so there are no model downloads and no local GPU to manage. + + + + +## What MiniMax H3 Image is good at + +- **2K and 4K output**: pick a size tier and an aspect ratio, or set an exact width and height +- **Image editing**: edit or combine up to 16 reference images, addressed in the prompt as `@Image1` to `@Image16` +- **Prompt rewriting**: MiniMax expands a short prompt into a detailed scene description by default and returns the rewritten prompt with the image +- **Poster-grade detail**: suited to posters, packaging, and editorial images that need legible text and dense detail + +## Text to image {#api_minimax_h3_image_t2i} + +Generate an image from a prompt at 2K or 4K. + +MiniMax H3 Image text to image workflow preview + + + + Download JSON or search "MiniMax-H3-Image: Text to Image (API)" in Template Library + + + +## Image edit {#api_minimax_h3_image_image_edit} + +Connect a source image, describe the change, and only what the prompt asks for is rewritten. The template adds a woman to a finished render and keeps the rest of the frame. + +MiniMax H3 Image edit workflow preview + + + + Download JSON or search "MiniMax-H3-Image: Image Edit (API)" in Template Library + + + +**Input materials** + + + + `LoadImage` · `surreal_rendering_modern.png` + + + +## Options + +| Input | Values | Notes | +| :---- | :----- | :---- | +| `model` | `MiniMax-H3-Image`, `MiniMax-H3-Image-Highspeed` | In the template, change the `model` dropdown to switch | +| `size` | `2K`, `4K`, `custom` | 2K and 4K derive width and height from the aspect ratio. `custom` takes a width and height from 512 to 7680 px each, with 655,360 to 20,000,000 px in total and an aspect ratio of at most 3:1 | +| `aspect_ratio` | `auto` plus 15 fixed ratios | `auto` follows the prompt, or the reference image when one is connected | +| `enhance_prompt` | on by default | MiniMax rewrites the prompt into a detailed scene description first. Edits are always enhanced, so turning it off only affects text to image, which then needs a fixed ratio or a custom size | +| `images` | up to 16 | Reference images for editing and combining. Refer to them in the prompt as `@Image1`, `@Image2`, and so on, in connection order | + +Prompts can be up to 8,000 characters when reference images are connected. The node also returns the rewritten `revised_prompt` next to the image. + +## Get started + +1. Update ComfyUI to 0.45.1 or later +2. In the Template Library, search for `MiniMax-H3-Image` diff --git a/zh/tutorials/partner-nodes/minimax/minimax-h3-image.mdx b/zh/tutorials/partner-nodes/minimax/minimax-h3-image.mdx new file mode 100644 index 0000000000..c5e8eb2838 --- /dev/null +++ b/zh/tutorials/partner-nodes/minimax/minimax-h3-image.mdx @@ -0,0 +1,79 @@ +--- +title: "MiniMax H3 Image API:在 ComfyUI 中进行文生图与图像编辑" +description: "通过 ComfyUI 合作伙伴 API 使用 MiniMax H3 Image,根据提示词生成 2K 或 4K 图像,或编辑最多 16 张参考图像。" +sidebarTitle: "MiniMax H3 Image" +translationSourceHash: ffae3038 +translationFrom: tutorials/partner-nodes/minimax/minimax-h3-image.mdx +translationBlockHashes: + "_intro": b2e9cda8 + "What MiniMax H3 Image is good at": 028a76d3 + "Text to image {#api_minimax_h3_image_t2i}": cd37d4b1 + "Image edit {#api_minimax_h3_image_image_edit}": cacb7043 + "Options": b178c63a + "Get started": d289ab9b +--- +import ReqHint from "/snippets/zh/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/zh/tutorials/update-reminder.mdx"; + +MiniMax H3 Image 是 MiniMax 的图像模型,可通过 **MiniMax H3 Image** 合作伙伴节点在 ComfyUI 中使用。文生图与图像编辑使用的是同一个节点:未连接参考图像时,它会根据你的提示词生成图像;连接了参考图像时,它会对这些图像进行编辑或组合。 + +这些工作流通过你的 Comfy API 账户在 MiniMax 的服务器上运行,因此无需下载模型,也无需管理本地 GPU。 + + + + +## MiniMax H3 Image 擅长什么 + +- **2K 和 4K 输出**:选择尺寸档位和宽高比,或设置精确的宽度和高度 +- **图像编辑**:最多可编辑或组合 16 张参考图像,在提示词中以 `@Image1` 至 `@Image16` 引用 +- **提示词重写**:MiniMax 默认会将简短提示词扩展为详细的场景描述,并随图像一并返回重写后的提示词 +- **海报级细节**:适合需要清晰文字和密集细节的海报、包装和编辑类图像 + +## 文本到图像 {#api_minimax_h3_image_t2i} + +根据提示词生成 2K 或 4K 图像。 + +MiniMax H3 Image 文本到图像工作流预览 + + + + 下载 JSON,或在模板库中搜索 "MiniMax-H3-Image: Text to Image (API)" + + + +## 图像编辑 {#api_minimax_h3_image_image_edit} + +连接一张来源图像并描述所需的修改,只有提示词要求的内容会被重写。该模板会在一张已完成的渲染图中添加一位女性,并保留画面其余部分。 + +MiniMax H3 图像编辑工作流预览 + + + + 下载 JSON,或在模板库中搜索 "MiniMax-H3-Image: Image Edit (API)" + + + +**输入素材** + + + + `LoadImage` · `surreal_rendering_modern.png` + + + +## 选项 + +| 输入 | 取值 | 说明 | +| :---- | :----- | :---- | +| `model` | `MiniMax-H3-Image`、`MiniMax-H3-Image-Highspeed` | 在模板中更改 `model` 下拉选项即可切换 | +| `size` | `2K`、`4K`、`custom` | 2K 和 4K 会根据宽高比推导出宽度和高度。`custom` 的宽度和高度各为 512 至 7680 px,总像素为 655,360 至 20,000,000 px,宽高比最大为 3:1 | +| `aspect_ratio` | `auto` 加上 15 种固定比例 | `auto` 会跟随提示词;若连接了参考图像,则跟随该图像 | +| `enhance_prompt` | 默认开启 | MiniMax 会先把提示词改写为详细的场景描述。编辑操作始终会执行增强,因此关闭它只影响文生图,此时需要选择固定比例或自定义尺寸 | +| `images` | 最多 16 张 | 用于编辑与合成的参考图像。在提示词中按连接顺序以 `@Image1`、`@Image2` 等引用它们 | + +连接参考图像时,提示词最多可达 8,000 个字符。节点还会在图像旁一并返回改写后的 `revised_prompt`。 + +## 开始使用 + +1. 将 ComfyUI 更新至 0.45.1 或更高版本 +2. 在模板库中搜索 `MiniMax-H3-Image`