爆炸檢視動畫
上傳裝置或物體的圖片以產生起始幀和結束幀。該工作流程將過渡動畫化為分解檢視並返回,輸出無縫迴圈的影片。
下載工作流程在 Cloud 中試用 API 文件
關於
這個 ComfyUI 工作流程「爆炸視圖動畫」可將裝置或物件的靜態圖像轉換為動態影片,展示爆炸視圖。流程中會使用 Nano Banana 2 與 Seedance 1.5 Pro 等模型來產生起始與結束畫格,並將其動畫化以形成無縫循環。關鍵節點如 ByteDanceFirstLastFrameNode 與 CreateVideo 負責畫格間的過渡,SaveImage 與 SaveVideo 節點則負責輸出。此流程特別適合用於可視化複雜機械結構或電子裝置,讓內部組件的展示更具吸引力,無需實體拆解。
您提供的內容
- 正在讀取工作流程
在 ComfyUI 中執行。 此工作流程無法在此頁面上執行。請在已載入所有內容的 Comfy Cloud 中開啟,或下載到您自己的 ComfyUI。
在 Cloud 中試用範例
工作流程內部
節點圖的唯讀預覽。請在 Comfy Cloud 中開啟,或下載後編輯節點。
Loading the graph
Open full-size workflow preview- 作者
- ComfyUI
需要自訂節點
- 正在讀取工作流程
從程式碼使用此工作流程
可行時會根據工作流程圖和輸入內容產生;否則會顯示儲存的範例。執行可能產生費用的程式碼前,請檢查認證資訊和端點。
1 取得 API 金鑰 ↗
What you need
- Runs on
- Comfy Cloud
- 金鑰
- COMFY_API_KEY
- 您的檔案
- 由程式碼透過其 URL 上傳
需要付費 Cloud 方案與點數。
下載 API JSONAPI 文件 ↗
# Python 3.10+: pip install comfy-sdk==0.4.0
import os
from comfy_sdk import Comfy
api_key = os.environ.get("COMFY_API_KEY", "").strip()
if not api_key: raise RuntimeError("Set COMFY_API_KEY")
os.environ.setdefault("COMFY_BASE_URL", "https://cloud.comfy.org")
client = Comfy(api_key=api_key)
workflow = client.workflows.from_json({
"26": {
"class_type": "SaveImage",
"inputs": {
"filename_prefix": "Nano Banana",
"images": [
"316",
0
]
},
"_meta": {
"title": "Save Image"
}
},
"29": {
"class_type": "LoadImage",
"inputs": {
"image": "UPLOADED_image_FILENAME"
},
"_meta": {
"title": "Load Image"
}
},
"225": {
"class_type": "SaveVideo",
"inputs": {
"filename_prefix": "video/ComfyUI-vid_1",
"format": "auto",
"format.codec": "auto",
"video": [
"244",
0
]
},
"_meta": {
"title": "Save Video"
}
},
"227": {
"class_type": "SaveVideo",
"inputs": {
"filename_prefix": "video/ComfyUI-vid_2",
"format": "auto",
"format.codec": "auto",
"video": [
"232",
0
]
},
"_meta": {
"title": "Save Video"
}
},
"232": {
"class_type": "ByteDanceFirstLastFrameNode",
"inputs": {
"aspect_ratio": "adaptive",
"camera_fixed": False,
"duration": 4,
"first_frame": [
"316",
0
],
"generate_audio": True,
"last_frame": [
"317",
0
],
"model": "seedance-1-5-pro-251215",
"prompt": "All expanded components slide back into place in a precise, mechanical motion.\n\nNo dialogue, narration, or voiceover — only visual movement.",
"resolution": "720p",
"seed": 642612484,
"watermark": False
},
"_meta": {
"title": "ByteDance First-Last-Frame to Video"
}
},
"241": {
"class_type": "SaveImage",
"inputs": {
"filename_prefix": "Nano Banana",
"images": [
"317",
0
]
},
"_meta": {
"title": "Save Image"
}
},
"244": {
"class_type": "ByteDanceFirstLastFrameNode",
"inputs": {
"aspect_ratio": "adaptive",
"camera_fixed": False,
"duration": 4,
"first_frame": [
"317",
0
],
"generate_audio": True,
"last_frame": [
"316",
0
],
"model": "seedance-1-5-pro-251215",
"prompt": "Slowly, its real physical components begin separating along natural assembly seams, transitioning into a clean, fully exploded view. Exterior casing, covers, buttons, ports, and internal structural parts separate in correct order and alignment, maintaining accurate scale and physical logic.\n\n No dialogue, no narration, no voiceover — only visual motion.",
"resolution": "720p",
"seed": 1561859130,
"watermark": False
},
"_meta": {
"title": "ByteDance First-Last-Frame to Video"
}
},
"259": {
"class_type": "BatchImagesNode",
"inputs": {
"images.image0": [
"392",
0
],
"images.image1": [
"393",
0
]
},
"_meta": {
"title": "Batch Images"
}
},
"316": {
"class_type": "GeminiNanoBanana2",
"inputs": {
"aspect_ratio": "auto",
"images": [
"317",
0
],
"model": "Nano Banana 2 (Gemini 3.1 Flash Image)",
"prompt": "a car",
"resolution": "2K",
"response_modalities": "IMAGE",
"seed": 343513052101185,
"system_prompt": "Analyze the reference image and recreate the exact same scene while transforming the main subject into a photorealistic exploded view.\n\n\nOnly the target object may change.\n\n\nOBJECT SEPARATION RULES\n\nThe object must be separated into its true physical components following realistic assembly structure.\n\nInternal components must be revealed and spaced apart along their correct assembly directions.\n\nAll parts must maintain accurate:\n\n- scale\n- thickness\n- orientation\n- alignment\n- relative position\n- physical hierarchy\n\nDo not invent components that would not physically exist.\n\nDo not duplicate identical components unless they exist multiple times in the real object.\n\n\nEXTERIOR COMPONENT PRESERVATION (CRITICAL)\n\n\nAll exterior parts must remain present and visible in the exploded view.\n\nThis includes:\n\n- front casing\n- back casing\n- side panels\n- outer shell\n- trim pieces\n- buttons\n- dials\n- knobs\n- ports\n- screens\n- external controls\n\nExterior casing must separate outward as real physical parts.\n\nThe shell must NOT disappear, dissolve, turn transparent, or be removed.\n\nThe result must NOT become an interior-only explosion.\n\n============================================================\nBACKGROUND AND CAMERA LOCK\n============================================================\n\nThe background must remain pixel-perfect identical to the original image.\n\nCamera angle, framing, and perspective must remain completely unchanged.\n\nDo NOT:\n\n- zoom out\n- zoom in\n- tilt\n- reframe\n- expand the canvas\n- change focal length\n\nObject scale within the frame must remain consistent with the original image.\n\n\nDo NOT include:\n\n- labels\n- callout graphics\n- diagrams\n- arrows\n- transparency\n- ghosted parts\n- stylization",
"thinking_level": "HIGH"
},
"_meta": {
"title": "Nano Banana 2"
}
},
"317": {
"class_type": "GeminiNanoBanana2",
"inputs": {
"aspect_ratio": "auto",
"images": [
"29",
0
],
"model": "Nano Banana 2 (Gemini 3.1 Flash Image)",
"prompt": "Analyze the input image and recreate it exactly as a single, static, photorealistic image.\n\nThis is an image recreation / image edit only.",
"resolution": "2K",
"response_modalities": "IMAGE",
"seed": 0,
"system_prompt": "You are an expert image-generation engine. You must ALWAYS produce an image.\nInterpret all user input—regardless of format, intent, or abstraction—as literal visual directives for image composition.\nIf a prompt is conversational or lacks specific visual details, you must creatively invent a concrete visual scenario that depicts the concept.\nPrioritize generating the visual representation above any text, formatting, or conversational requests.",
"thinking_level": "MINIMAL"
},
"_meta": {
"title": "Nano Banana 2"
}
},
"380": {
"class_type": "CreateVideo",
"inputs": {
"audio": [
"392",
1
],
"bit_depth": "auto",
"codec": "none",
"color_space": "sRGB",
"fps": [
"392",
2
],
"images": [
"259",
0
]
},
"_meta": {
"title": "Create Video"
}
},
"381": {
"class_type": "SaveVideo",
"inputs": {
"filename_prefix": "video/ComfyUI",
"format": "auto",
"format.codec": "auto",
"video": [
"380",
0
]
},
"_meta": {
"title": "Save Video"
}
},
"392": {
"class_type": "GetVideoComponents",
"inputs": {
"video": [
"244",
0
]
},
"_meta": {
"title": "Get Video Components"
}
},
"393": {
"class_type": "GetVideoComponents",
"inputs": {
"video": [
"232",
0
]
},
"_meta": {
"title": "Get Video Components"
}
}
})
# Replace each example.com URL with a public URL of your file.
workflow.set_input("29", "image", client.assets.from_url("https://example.com/replace-with-a-url/image"))
# run() also passes api_key to the partner nodes in this workflow.
job = client.run(workflow, api_key=api_key)問答
本工作流程不支援音訊。你可以在影片匯出後,使用影片剪輯工具加入音訊。
具有清晰輪廓與細節的裝置或物件圖片效果最佳,若有內部結構的參考圖會更好。
啟用內部組件參考圖的節點,能提升準確度,特別是在需要精確內部結構時。
可以,你可以在合成影片前,利用 SaveImage 節點儲存單張畫格。