[
  {
    "id": "001",
    "title": "Pixal3D 与 TRELLIS.2 工作流",
    "description": "P1  ·  3D 资产。图片到模型，供预演场景资产准备；网格、材质和导出质量需实测。",
    "links": [
      "https://comfy.org/workflows/920c795f693a-920c795f693a/"
    ],
    "status": "已下载",
    "note": "原工作流页面的预览，不代表已在本机安装或运行。",
    "leads": [],
    "videos": [
      {
        "id": "001",
        "label": "Pixal3D与TRELLIS工作流预览",
        "url": "https://media.comfy.org/hub-media/video/69a3a477-f1e0-4ea6-b09c-12ad80b577db.mp4",
        "source": "https://comfy.org/workflows/920c795f693a-920c795f693a/",
        "kind": "原条目演示",
        "source_size": 1649829,
        "duration": 8.85,
        "width": 870,
        "height": 870,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/001_Pixal3D与TRELLIS工作流预览.mp4",
        "bytes": 1649829,
        "sha256": "9ba309bc062df7747bf8ca5acb9de6f0678d68ccd15c0df72ffde54ae893c196"
      }
    ]
  },
  {
    "id": "002",
    "title": "video2dlssnr",
    "description": "P2  ·  画质处理。视频超分与画质增强参考；Windows RTX 路线，不能当作补帧。",
    "links": [
      "https://github.com/DaniilSokolyuk/video2dlssnr"
    ],
    "status": "未下载",
    "note": "已查原仓库与项目名搜索，本轮未找到可下载的独立短演示。原仓库以说明和界面截图为主。",
    "leads": [],
    "videos": []
  },
  {
    "id": "003",
    "title": "Seedance 2.5 Video Director",
    "description": "P0  ·  创作规则。导演方案、时间线提示词、参考用途与模型适配；属于提示词 Skill。",
    "links": [
      "https://github.com/liyue-aigc/seedance-2-5-video-director"
    ],
    "status": "未下载",
    "note": "原仓库是导演提示词Skill。本轮未找到能明确归属于该Skill的公开短演示，不用通用Seedance样片代替。",
    "leads": [],
    "videos": []
  },
  {
    "id": "004",
    "title": "LTX Ripple Demo",
    "description": "P0 验证  ·  视频编辑。修改首帧后传播至短视频。Demo 有片长、尺寸与帧率处理约束，声音需单独管理。",
    "links": [
      "https://huggingface.co/spaces/WepeNerd/ltx-ripple-demo"
    ],
    "status": "已下载",
    "note": "从原Space链接到Ripple模型作者页面，下载其效果示例；不是Space界面录屏。",
    "leads": [],
    "videos": [
      {
        "id": "004",
        "label": "Ripple首帧修改传播演示",
        "url": "https://huggingface.co/WepeNerd/LTX-Ripple/resolve/main/Examples/ltx_001.mp4",
        "source": "https://huggingface.co/WepeNerd/LTX-Ripple",
        "kind": "作者模型效果演示",
        "note": "来自原Space所使用的Ripple模型作者页面",
        "source_size": 7810326,
        "duration": 4.84,
        "width": 1152,
        "height": 1536,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/004_Ripple首帧修改传播演示.mp4",
        "bytes": 7810326,
        "sha256": "f235f8abbb1d3acb55caec3d95818dd604ddb5e228facea8b0aad0c6d82664a5"
      }
    ]
  },
  {
    "id": "005",
    "title": "SAM 3D Body 到 Blender",
    "description": "P1  ·  动作预演。用户补充推文链路为录像提动作、Blender 改机位、交给 Seedance；Cloud 图未完整核验。",
    "links": [
      "https://x.com/ComfyUI/status/2094457017620693232",
      "https://cloud.comfy.org/?template=utility_sam3d_body",
      "https://cloud.comfy.org/?share=a51863b6a761"
    ],
    "status": "已下载",
    "note": "下载的是附录原ComfyUI推文视频，包含动作参考、Blender机位与生成结果。",
    "leads": [],
    "videos": [
      {
        "id": "005",
        "label": "SAM动作提取到Blender到Seedance",
        "url": "https://video.twimg.com/amplify_video/2094456962243231744/vid/avc1/720x1280/PaGpWTNG8UgVRWIq.mp4?tag=14",
        "source": "https://x.com/ComfyUI/status/2094457017620693232",
        "kind": "原推文演示",
        "source_size": 2284367,
        "duration": 16.106667,
        "width": 720,
        "height": 1280,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/005_SAM动作提取到Blender到Seedance.mp4",
        "bytes": 2284367,
        "sha256": "e4b8dfdc5f9e92a34eeadb6b0da5f72c3b99c632a86282af2db10c4f11de8a62"
      }
    ]
  },
  {
    "id": "006",
    "title": "H3 Storyboard Skill",
    "description": "P0  ·  分镜规则。节拍、表演、对白与镜头组织，适合分镜编译；不将对白视为表情动画的必要条件。",
    "links": [
      "https://github.com/phileiny/h3-storyboard-skill"
    ],
    "status": "未下载",
    "note": "已查原仓库与公开介绍，发现分镜规则和对照实验说明，但本轮未找到可下载的独立演示视频。",
    "leads": [],
    "videos": []
  },
  {
    "id": "007",
    "title": "MiniMax H3 Acc LoRAs",
    "description": "P1  ·  快速预览。FL2VA／Ref2VA 加速候选。少步数不等于同比例加速，需比较草稿可用率。",
    "links": [
      "https://huggingface.co/alibaba-pai/MiniMax-H3-Acc-LoRAs"
    ],
    "status": "已下载",
    "note": "保留基线和4步加速各一段便于比较画面；播放时长不是推理耗时。",
    "leads": [],
    "videos": [
      {
        "id": "007",
        "label": "H3原始步数基线样片",
        "url": "https://huggingface.co/alibaba-pai/MiniMax-H3-Acc-LoRAs/resolve/main/results/minimax_h3_fl2va_baseline_1.mp4",
        "source": "https://huggingface.co/alibaba-pai/MiniMax-H3-Acc-LoRAs",
        "kind": "作者效果样片",
        "source_size": 835589,
        "duration": 5.175,
        "width": 1344,
        "height": 768,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/007_H3原始步数基线样片.mp4",
        "bytes": 835589,
        "sha256": "afcbd76e7e5b6107df3b5b0a77fbdc09253d2e20beca55403730d45c55032584"
      },
      {
        "id": "007",
        "label": "H3四步加速对照样片",
        "url": "https://huggingface.co/alibaba-pai/MiniMax-H3-Acc-LoRAs/resolve/main/results/minimax_h3_fl2va_turbo_4step_v1.1_768p_1.mp4",
        "source": "https://huggingface.co/alibaba-pai/MiniMax-H3-Acc-LoRAs",
        "kind": "作者效果样片",
        "note": "与同编号基线并列查看，不以视频播放时长代表推理耗时",
        "source_size": 893312,
        "duration": 5.175,
        "width": 1344,
        "height": 768,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/007_H3四步加速对照样片.mp4",
        "bytes": 893312,
        "sha256": "b92514db1f7a59517a1ef165748e5d7fab6e317fc51e8212e1b95645e3912f18"
      }
    ]
  },
  {
    "id": "008",
    "title": "Reference Video Director",
    "description": "P0  ·  连续性规则。围绕参考图片组织身份、视点、互动和跨段状态；不能按名称认定支持视频解析。",
    "links": [
      "https://github.com/sgyno09-source/reference-video-director"
    ],
    "status": "未下载",
    "note": "原条目为参考图片导演Skill，未找到明确对应的公开演示短片。",
    "leads": [],
    "videos": []
  },
  {
    "id": "009",
    "title": "4DAnyone",
    "description": "P2  ·  多视图研究。从单目人物视频生成多视图；下游 3D／4D 重建另行验证，不能保证直接获得可编辑角色。",
    "links": [
      "https://github.com/ant-research/4DAnyone/tree/main"
    ],
    "status": "已下载",
    "note": "作者原始GIF转为MP4便于播放；这是多视角效果动画，不是可编辑3D资产。",
    "leads": [],
    "videos": [
      {
        "id": "009",
        "label": "4DAnyone多视角概览",
        "url": "https://raw.githubusercontent.com/ant-research/4DAnyone/HEAD/docs/assets/teaser.gif",
        "source": "https://github.com/ant-research/4DAnyone",
        "kind": "作者动画演示",
        "source_size": 15923562,
        "conversion": "作者原始动画GIF转MP4；未新增画面；无音轨",
        "duration": 6.0,
        "width": 960,
        "height": 540,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/009_4DAnyone多视角概览.mp4",
        "bytes": 1491210,
        "sha256": "55004923326f8b5b4755a14c3fb3629153ad05346917a0fa32701c82fbe89b92"
      }
    ]
  },
  {
    "id": "010",
    "title": "shuohao skills",
    "description": "P0  ·  故事与设定。小说改编、分集、人物、美术、剧本等规则；长片场景需调整原有短剧模板。",
    "links": [
      "https://github.com/eternityspring/shuohao-skills"
    ],
    "status": "未下载",
    "note": "找到33秒的社区项目介绍，但YouTube要求登录验证，未下载。未读取浏览器账号或Cookie。",
    "leads": [
      {
        "title": "33秒社区介绍：Turn Any Book Into Visual Concepts With Shuohao Skills",
        "url": "https://www.youtube.com/watch?v=0VXzkCE_j-Q"
      }
    ],
    "videos": []
  },
  {
    "id": "011",
    "title": "写真提示词规划",
    "description": "P2  ·  美术设定。参考图分析、系列母题与画面变体；保留为角色造型图辅助。",
    "links": [
      "https://github.com/nuyoah-ai-works/nuyoah-xiezhen-prompt"
    ],
    "status": "未下载",
    "note": "原仓库及作者说明页以写真提示词和静态图为主，本轮未找到可下载的独立短演示。",
    "leads": [
      {
        "title": "作者说明页",
        "url": "https://knowledge.nuyoahonline.com/skills/nuyoah-xiezhen-prompt"
      }
    ],
    "videos": []
  },
  {
    "id": "012",
    "title": "img2threejs",
    "description": "P1  ·  程序化场景。由图片组织 Three.js 场景和部件编辑；代码场景与可导出的标准网格需区分。",
    "links": [
      "https://github.com/img2threejs/img2threejs"
    ],
    "status": "已下载",
    "note": "项目讨论区引用的社区演示。发布者称使用img2threejs与Codex，未在本机复现生成过程。",
    "leads": [],
    "videos": [
      {
        "id": "012",
        "label": "img2threejs社区建模效果演示",
        "url": "https://video.twimg.com/amplify_video/2078996632574496768/vid/avc1/816x720/w1z74IKCGqsebG6w.mp4?tag=29",
        "source": "https://x.com/hamster_wat/status/2078996887361794149",
        "kind": "社区使用演示",
        "note": "作者项目讨论区引用的社区案例；发布者称使用img2threejs与Codex，本地未复现其生成过程",
        "source_size": 3246191,
        "duration": 43.221333,
        "width": 816,
        "height": 720,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/012_img2threejs社区建模效果演示.mp4",
        "bytes": 3246191,
        "sha256": "c6cba41acc382ebd934c7d2aa3b3a2e0cb93f5c80c8f20a78770cc2ba14b7dfd"
      }
    ]
  },
  {
    "id": "013",
    "title": "Seedream 图层分离",
    "description": "P1  ·  图像编辑。图层拆解与合成，可用于修改分镜／首帧；精确蒙版和文件交付格式需验证。",
    "links": [
      "https://cloud.comfy.org/?template=api_bytedance_seedream_5_0_layer_separation"
    ],
    "status": "已下载",
    "note": "Comfy官方成员公开发布的图层分离与PSD交付演示。",
    "leads": [],
    "videos": [
      {
        "id": "013",
        "label": "Seedream图层分离与PSD导出",
        "url": "https://dms.licdn.com/playlist/vid/v2/D5610AQHnxKvDaw9ZOg/mp4-720p-30fp-crf28/B56aAbOqUNIsBk-/0/1787163231849?e=2147483647&v=beta&t=kQxICd8HEeS84fmi_7uBZ9YLq37yjsvt2Q9MYWfzBOc",
        "source": "https://www.linkedin.com/posts/jo-zhang-178b27103_comfyui-generativeai-activity-7495905894422011904-C5tL",
        "kind": "官方成员操作演示",
        "source_size": 1836393,
        "duration": 39.4,
        "width": 1280,
        "height": 720,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/013_Seedream图层分离与PSD导出.mp4",
        "bytes": 1836393,
        "sha256": "32a6f8ca063dcf8eb9c4ec307a4315badeb526565fc35dc4bb60d3002da5b7b5"
      }
    ]
  },
  {
    "id": "014",
    "title": "SCOPE Camera Video Generation",
    "description": "P1  ·  受控运镜。首帧与显式相机路径生成候选；先验证支持路径、实际帧数和可控范围。",
    "links": [
      "https://huggingface.co/spaces/TencentARC/scope-camera-video-generation"
    ],
    "status": "已下载",
    "note": "保留短运镜样例和森林对照动画。另找到的项目概览120.2秒，按2分钟上限跳过。",
    "leads": [],
    "videos": [
      {
        "id": "014",
        "label": "SCoPE复杂运镜与轨迹对照",
        "url": "https://huggingface.co/datasets/Michaelqaz/SCoPE-Gallery/resolve/main/videos/airmountains__env_scenic_grand_tour_22_page_pip.mp4",
        "source": "https://visual-ai.github.io/scope/",
        "kind": "官方项目效果演示",
        "source_size": 5086248,
        "duration": 5.0625,
        "width": 832,
        "height": 480,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/014_SCoPE复杂运镜与轨迹对照.mp4",
        "bytes": 5086248,
        "sha256": "c98e148bee41e830cfb1ccb1424f350fa9bc5928b1490e657a67931797cf1a7c"
      },
      {
        "id": "014",
        "label": "SCoPE森林右移对照",
        "url": "https://raw.githubusercontent.com/TencentARC/SCoPE/HEAD/assets/demo/forest_truck_right.gif",
        "source": "https://github.com/TencentARC/SCoPE",
        "kind": "作者动画演示",
        "source_size": 1432788,
        "conversion": "作者原始动画GIF转MP4；未新增画面；无音轨",
        "duration": 5.125,
        "width": 320,
        "height": 186,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/014_SCoPE森林右移对照.mp4",
        "bytes": 115773,
        "sha256": "f62ad917ac5b8fbe002743f7447af8546b43ad3d4df9f4d40bd30b7116c57499"
      }
    ]
  },
  {
    "id": "015",
    "title": "Open Magiviz",
    "description": "P0  ·  制作编排。任务、进度、局部重做、批次及粗剪的流程参考；README 能力尚未运行验证。",
    "links": [
      "https://github.com/ItusiAI/Open-Magiviz"
    ],
    "status": "未下载",
    "note": "已查原仓库、官方站和公开页面资源，本轮未取得明确对应且符合大小限制的演示短片。",
    "leads": [],
    "videos": []
  },
  {
    "id": "016",
    "title": "H3 Realism People LoRA",
    "description": "P2  ·  人物风格。写实人物视觉预设；与加速或身份 LoRA 的组合另行验证。",
    "links": [
      "https://huggingface.co/fal/MiniMax-H3-Realism-People-LoRA"
    ],
    "status": "已下载",
    "note": "Hugging Face的对比源文件约43MiB，已跳过；改用模型作者页面所列的较小fal外链版本。",
    "leads": [],
    "videos": [
      {
        "id": "016",
        "label": "H3写实人物作者外链对比",
        "url": "https://v3b.fal.media/files/b/0aa5ce26/stD_vdWuX1F90Q2OihK1Q_learn_before_after.mp4",
        "source": "https://huggingface.co/fal/MiniMax-H3-Realism-People-LoRA",
        "kind": "作者对比演示",
        "source_size": 4594479,
        "duration": 5.0,
        "width": 1920,
        "height": 540,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/016_H3写实人物作者外链对比.mp4",
        "bytes": 4594479,
        "sha256": "c97683cd3999cebd522c4c2c5e46878ea77364169e3205cdc1ddd860c2279304"
      }
    ]
  },
  {
    "id": "017",
    "title": "Seedance TVC Director",
    "description": "P2  ·  广告创作。品牌简报、产品动作、多时长广告和后期交接，作为长片之外的扩展。",
    "links": [
      "https://github.com/nutllwhy/seedance-tvc-director"
    ],
    "status": "未下载",
    "note": "仓库主要提供导演Skill规则。搜索发现作者TVC样片发布线索，但本轮未取得可核验的直接视频地址，未用其他广告素材冒充。",
    "leads": [],
    "videos": []
  },
  {
    "id": "018",
    "title": "Character Swap V2V Prompts",
    "description": "待核验  ·  视频换角。Notion 正文未能读取，暂不形成已确认功能；后续获取内容再评估。",
    "links": [
      "https://app.notion.com/p/CHARACTER-SWAP-V2V-PROMPTS-halo-christo-3b663c2a7548802a8143d3ac8cb23373"
    ],
    "status": "未下载",
    "note": "原Notion链接未能核验正文，无法可靠确认对应演示。不以同名换角视频冒充。",
    "leads": [],
    "videos": []
  },
  {
    "id": "019",
    "title": "42 Camera Movements",
    "description": "P1  ·  运镜词库。用于可检索镜头语言与模板；相机运动、光学效果和时间效果分开建模。",
    "links": [
      "https://aishotstudio.com/42-camera-movements-ai-prompts/"
    ],
    "status": "已下载",
    "note": "从本条运镜库选择推进、拉远、过肩、横移、环绕、升高6个代表示例。它们是同一个019条目的子例，不是附录另外6条。",
    "leads": [],
    "videos": [
      {
        "id": "019",
        "label": "01缓慢推进",
        "url": "https://aishotstudio.com/wp-content/uploads/2026/01/1.-Slow-Dolly-In_preview.mp4",
        "source": "https://aishotstudio.com/42-camera-movements-ai-prompts/",
        "kind": "原条目运镜示例",
        "source_size": 452303,
        "duration": 4.550998,
        "width": 848,
        "height": 480,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/019_01缓慢推进.mp4",
        "bytes": 452303,
        "sha256": "0a9ec97cc894c94f06c359da6d6d5d215a19b4985decf1c675fa068a84e09f95"
      },
      {
        "id": "019",
        "label": "02缓慢拉远",
        "url": "https://aishotstudio.com/wp-content/uploads/2026/01/2.-Slow-Dolly-Out_preview.mp4",
        "source": "https://aishotstudio.com/42-camera-movements-ai-prompts/",
        "kind": "原条目运镜示例",
        "source_size": 553619,
        "duration": 4.852993,
        "width": 848,
        "height": 480,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/019_02缓慢拉远.mp4",
        "bytes": 553619,
        "sha256": "5888e45d187f0941d5b3b1f26ef92d111be5ea3c760f4ab52dfd1f46b09954f5"
      },
      {
        "id": "019",
        "label": "03过肩镜头",
        "url": "https://aishotstudio.com/wp-content/uploads/2026/01/7.-Over-the-Shoulder-OTS-Shot_preview.mp4",
        "source": "https://aishotstudio.com/42-camera-movements-ai-prompts/",
        "kind": "原条目镜头示例",
        "source_size": 513858,
        "duration": 4.852993,
        "width": 848,
        "height": 480,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/019_03过肩镜头.mp4",
        "bytes": 513858,
        "sha256": "540edddc2c497ca6d6a31d255b5cd5deac6a3a165314b028aae4ebe7f124bc1e"
      },
      {
        "id": "019",
        "label": "04左向横移",
        "url": "https://aishotstudio.com/wp-content/uploads/2026/01/15.-Camera-Truck-Left_preview.mp4",
        "source": "https://aishotstudio.com/42-camera-movements-ai-prompts/",
        "kind": "原条目运镜示例",
        "source_size": 607829,
        "duration": 4.852993,
        "width": 848,
        "height": 480,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/019_04左向横移.mp4",
        "bytes": 607829,
        "sha256": "cbbc2a31dc00b9728cda82620efad80b9919180746656c85d445605f2262da70"
      },
      {
        "id": "019",
        "label": "05半圈环绕",
        "url": "https://aishotstudio.com/wp-content/uploads/2026/01/True-Orbit-180-preview.mp4",
        "source": "https://aishotstudio.com/42-camera-movements-ai-prompts/",
        "kind": "原条目运镜示例",
        "source_size": 808689,
        "duration": 7.083333,
        "width": 854,
        "height": 480,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/019_05半圈环绕.mp4",
        "bytes": 808689,
        "sha256": "c74bd2fc400203c2ff3aaf8c23d41d069bb001b292694e1c7b92e8e42da62af7"
      },
      {
        "id": "019",
        "label": "06摇臂升高",
        "url": "https://aishotstudio.com/wp-content/uploads/2026/01/22.-Crane-Up-High-Angle-Reveal_preview.mp4",
        "source": "https://aishotstudio.com/42-camera-movements-ai-prompts/",
        "kind": "原条目运镜示例",
        "source_size": 573528,
        "duration": 4.852993,
        "width": 848,
        "height": 480,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/019_06摇臂升高.mp4",
        "bytes": 573528,
        "sha256": "223cb4407756ec91aa926d0bde450149ec1f46071b449b3940ed45f22891d979"
      }
    ]
  },
  {
    "id": "020",
    "title": "LTX 2.3 IC LoRA Relight",
    "description": "P1  ·  视频重打光。外景太阳方向控制，使用光照球参考；不承诺室内多灯或任意光线控制。",
    "links": [
      "https://huggingface.co/Lightricks/LTX-2.3-22b-IC-LoRA-Relight"
    ],
    "status": "未下载",
    "note": "找到官方重打光前后对比，但原文件要求Hugging Face登录并同意访问条件；本轮未下载。公开Space只有可直接读取的输入素材，不把输入视频当成重打光结果。",
    "leads": [
      {
        "title": "官方重打光在线演示",
        "url": "https://huggingface.co/spaces/Lightricks/ltx2-relight-lora-demo"
      }
    ],
    "videos": []
  },
  {
    "id": "021",
    "title": "awplanet",
    "description": "P0 参考  ·  虚拟片场。场景摆放、人物姿态、摄影机及时间线参考；商业代码采用须先核对授权。",
    "links": [
      "https://github.com/awplanets/awplanet"
    ],
    "status": "已下载",
    "note": "经社区转帖找到作者DynamicWang的15秒原始媒体，保留作者推文来源。",
    "leads": [],
    "videos": [
      {
        "id": "021",
        "label": "awplanet手机虚拟运镜演示",
        "url": "https://video.twimg.com/amplify_video/2072947745556717568/vid/avc1/720x1280/N1wWaHL0U_gHMR_i.mp4?tag=28",
        "source": "https://x.com/dynamicwangs/status/2072948527949582658",
        "kind": "作者原推文演示",
        "note": "经社区转帖找到，媒体发布者为AWPlanet创始人DynamicWang",
        "source_size": 1275359,
        "duration": 15.11619,
        "width": 720,
        "height": 1280,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/021_awplanet手机虚拟运镜演示.mp4",
        "bytes": 1275359,
        "sha256": "16947ca3e5c0d1226b4f4070eacf4298a2029e5d22d8cf8009fc6f4f07ceb0a8"
      }
    ]
  },
  {
    "id": "022",
    "title": "LTX Black Magic LoRA",
    "description": "P2  ·  暗部重建。暗画面生成式重建与增强；不能视为忠实恢复隐藏信息。",
    "links": [
      "https://huggingface.co/FuzzPuppy/LTX-2.3-Black-Magic-LoRA"
    ],
    "status": "已下载",
    "note": "作者效果样片；生成式暗部重建不等于真实信息恢复。",
    "leads": [],
    "videos": [
      {
        "id": "022",
        "label": "暗部重建双人舞样片",
        "url": "https://huggingface.co/FuzzPuppy/LTX-2.3-Black-Magic-LoRA/resolve/main/examples/dance-couple.mp4",
        "source": "https://huggingface.co/FuzzPuppy/LTX-2.3-Black-Magic-LoRA",
        "kind": "作者效果演示",
        "source_size": 5725351,
        "duration": 5.039002,
        "width": 1024,
        "height": 1140,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/022_暗部重建双人舞样片.mp4",
        "bytes": 5725351,
        "sha256": "87e2c02e1282897623c0da32c379faa5a310d6f9e179db91d80d749d89ecbac0"
      }
    ]
  },
  {
    "id": "023",
    "title": "Comfy 换脸工作流",
    "description": "P1  ·  视频换脸。参考图驱动的视频换脸／年龄修改候选；时序、遮挡和表演保留需测试。",
    "links": [
      "https://comfy.org/workflows/93f286fbc2c8-93f286fbc2c8/"
    ],
    "status": "未下载",
    "note": "已找到原工作流介绍视频，但时长249.8秒，超过2分钟限制，已跳过，未下载完整长片。",
    "leads": [
      {
        "title": "原工作流介绍视频，249.8秒，仅保留链接",
        "url": "https://media.comfy.org/hub-media/video/d0bc92f7-c5dc-4861-9588-5ae94b609c6d.mp4"
      }
    ],
    "videos": []
  },
  {
    "id": "024",
    "title": "LTX Lens Remover",
    "description": "P2  ·  光学清理。清理光学伪影及相关外观；有生成式修改风险，Log 输出与源色彩需确认。",
    "links": [
      "https://huggingface.co/JanKanta/ltx-2.3-22b-ic-lora-lens_remover"
    ],
    "status": "已下载",
    "note": "完整版Showcase约42MiB，已跳过；仅保留作者较小的sRGB结果样片，不是完整前后对比。",
    "leads": [],
    "videos": [
      {
        "id": "024",
        "label": "LensRemover清理后sRGB样片",
        "url": "https://huggingface.co/JanKanta/ltx-2.3-22b-ic-lora-lens_remover/resolve/main/srgb.mp4",
        "source": "https://huggingface.co/JanKanta/ltx-2.3-22b-ic-lora-lens_remover",
        "kind": "作者效果样片",
        "source_size": 3835476,
        "duration": 4.84,
        "width": 1920,
        "height": 540,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/024_LensRemover清理后sRGB样片.mp4",
        "bytes": 3835476,
        "sha256": "25ce25ef26d0995481e2bfdc4cab345b7734cc4a614ebcf708bff3075600bc9b"
      }
    ]
  },
  {
    "id": "025",
    "title": "Comfy 分享工作流",
    "description": "待核验  ·  用途未明。分享页面未读取到有效内容，需标题、截图或工作流 JSON 后归类。",
    "links": [
      "https://cloud.comfy.org/?share=135f7596d783"
    ],
    "status": "未下载",
    "note": "原Comfy分享链接未读取到可确认的工作流名称与内容，无法匹配演示。",
    "leads": [],
    "videos": []
  },
  {
    "id": "026",
    "title": "Krea2 ReID",
    "description": "P0 验证  ·  图片身份参考。以单张 RGB 参考生成身份相关图像变体，适合首帧准备；不等于视频换脸。",
    "links": [
      "https://huggingface.co/yijunwang2/krea2-reid"
    ],
    "status": "未下载",
    "note": "模型页及文件目录主要是静态身份参考图，未找到独立演示视频。不把静态WebP图转成伪视频充数。",
    "leads": [],
    "videos": []
  },
  {
    "id": "027",
    "title": "Blender MCP",
    "description": "P0 参考  ·  3D 连接层。通过助手操作 Blender，需另建 Narra 场景映射、保存回传和操作恢复。",
    "links": [
      "https://github.com/ahujasid/blender-mcp"
    ],
    "status": "已下载",
    "note": "YouTube提取遇到登录验证，改为下载项目作者X上公开发布的22.74秒演示。",
    "leads": [],
    "videos": [
      {
        "id": "027",
        "label": "BlenderMCP自然语言创建场景",
        "url": "https://video.twimg.com/ext_tw_video/1899457044270465024/pu/vid/avc1/1280x720/Gfcg-3ef4tKe5Tob.mp4?tag=14",
        "source": "https://x.com/sidahuj/status/1899460492999184534",
        "kind": "作者原推文演示",
        "source_size": 4173925,
        "duration": 22.784,
        "width": 1280,
        "height": 720,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/027_BlenderMCP自然语言创建场景.mp4",
        "bytes": 4173925,
        "sha256": "8f35bd9496a56fb6dea0530aa73883daf03067d2fe8fb7af8f12517c5a494fa7"
      }
    ]
  },
  {
    "id": "028",
    "title": "DiffusionGemma Prompt Builder",
    "description": "P0  ·  编译与编排。多模态上下文、模型专用任务包、校验、音频与分段制作；核心不执行视频采样。",
    "links": [
      "https://github.com/exportAnything/ComfyUI-DiffusionGemmaPromptBuilder"
    ],
    "status": "已下载",
    "note": "作者GIF展示完整任务链路；Prompt Builder本身不是视频采样器。",
    "leads": [],
    "videos": [
      {
        "id": "028",
        "label": "任务编排角色动作迁移演示",
        "url": "https://raw.githubusercontent.com/exportAnything/ComfyUI-DiffusionGemmaPromptBuilder/HEAD/examples/assets/character_motion_transfer/diffusiongemma_ltx_character_demo.gif",
        "source": "https://github.com/exportAnything/ComfyUI-DiffusionGemmaPromptBuilder",
        "kind": "作者动画演示",
        "note": "显示完整链路样例，Prompt Builder本身不执行视频采样",
        "source_size": 9192018,
        "conversion": "作者原始动画GIF转MP4；未新增画面；无音轨",
        "duration": 29.375,
        "width": 640,
        "height": 360,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/028_任务编排角色动作迁移演示.mp4",
        "bytes": 848811,
        "sha256": "c42bcd5b62114ae3916c6b2c9f944dee9fdeb921d601356365ec9b0349b3dfa0"
      }
    ]
  },
  {
    "id": "029",
    "title": "LTX CrossView Prompt",
    "description": "P1  ·  替代机位。视频新视角概念验证；离散机位词汇、小幅变更优先，背面不在训练范围。",
    "links": [
      "https://huggingface.co/Cseti/LTX2.3-22B_IC-LoRA-CrossView-Prompt"
    ],
    "status": "已下载",
    "note": "作者替代机位对比样例，用于观看视角变化与连续性，不代表任意角度都可控。",
    "leads": [],
    "videos": [
      {
        "id": "029",
        "label": "CrossView替代机位对比",
        "url": "https://huggingface.co/Cseti/LTX2.3-22B_IC-LoRA-CrossView-Prompt/resolve/main/assets/example_warrior_right-lower-closer.mp4",
        "source": "https://huggingface.co/Cseti/LTX2.3-22B_IC-LoRA-CrossView-Prompt",
        "kind": "作者对比演示",
        "source_size": 3267443,
        "duration": 3.708333,
        "width": 960,
        "height": 1132,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/029_CrossView替代机位对比.mp4",
        "bytes": 3267443,
        "sha256": "c67decbc5e035141db60ef2b866b998fb22cf4d44a744336e17111ef1c40fa50"
      }
    ]
  },
  {
    "id": "030",
    "title": "LingBot Video",
    "description": "P1  ·  动作生成。动作交互候选引擎，结构化提示词与 Refiner；具身评测不等于长片表演质量。",
    "links": [
      "https://github.com/robbyant/lingbot-video"
    ],
    "status": "已下载",
    "note": "原仓库作者发布的动作交互演示。",
    "leads": [],
    "videos": [
      {
        "id": "030",
        "label": "LingBot动作交互演示",
        "url": "https://github.com/user-attachments/assets/e960aa83-6e5c-42ee-ba33-9fa25749ac03",
        "source": "https://github.com/robbyant/lingbot-video",
        "kind": "作者演示",
        "source_size": 9682507,
        "duration": 62.156544,
        "width": 720,
        "height": 576,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/030_LingBot动作交互演示.mp4",
        "bytes": 9682507,
        "sha256": "bdc98a1d019f75ec00b73399a423631b7e4de9088fe661f8b11d2f9c68c4a5a3"
      }
    ]
  },
  {
    "id": "031",
    "title": "PhotoGIMP",
    "description": "P2  ·  编辑交互。GIMP 界面与快捷键配置补丁，供图像编辑器交互参考，不是网页编辑引擎。",
    "links": [
      "https://github.com/Diolinux/Photogimp"
    ],
    "status": "未下载",
    "note": "已查原仓库及作者教程线索，本轮未取得符合长度、大小及公开访问要求的短片；保留原项目链接。PhotoGIMP是界面配置补丁。",
    "leads": [
      {
        "title": "作者PhotoGIMP安装介绍",
        "url": "https://plus.diolinux.com.br/t/como-instalar-o-photogimp-para-deixar-o-gimp-mais-parecido-com-o-photoshop/70738"
      }
    ],
    "videos": []
  },
  {
    "id": "032",
    "title": "LTX Preproduction Playground",
    "description": "待核验  ·  预制作工作台。Space 页面及源码未能读取。与 Previz 高度相关，但不据名称确认具体能力。",
    "links": [
      "https://huggingface.co/spaces/ltx-community/ltx-2-3-preproduction-playground"
    ],
    "status": "未下载",
    "note": "本轮已读到Space公开说明与源码，较原文档新增了可读资料；未找到独立可下载的工作台操作短片。对应X链接是预制作文章。",
    "leads": [
      {
        "title": "LTX预制作原文",
        "url": "https://x.com/ltx_io/status/2074169403764084834"
      }
    ],
    "videos": []
  },
  {
    "id": "033",
    "title": "LTX Best Face ID",
    "description": "P0 验证  ·  角色视频参考。人脸版与四视图角色表版；后者要求 1536×1024 参考图。不是直接成片换脸。",
    "links": [
      "https://huggingface.co/Alissonerdx/LTX-Best-Face-ID"
    ],
    "status": "已下载",
    "note": "已保留人脸身份参考样片；另一个四视图角色表候选地址返回404，未纳入。",
    "leads": [],
    "videos": [
      {
        "id": "033",
        "label": "FaceID人物身份参考样片",
        "url": "https://huggingface.co/Alissonerdx/LTX-Best-Face-ID/resolve/main/examples/sample_1.mp4",
        "source": "https://huggingface.co/Alissonerdx/LTX-Best-Face-ID",
        "kind": "作者效果样片",
        "source_size": 3187643,
        "duration": 5.041667,
        "width": 1838,
        "height": 1168,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/033_FaceID人物身份参考样片.mp4",
        "bytes": 3187643,
        "sha256": "0ee90e7def91114800dd9deb24140f759918430c3144ba5844e349cc3e6b5d7d"
      }
    ]
  },
  {
    "id": "034",
    "title": "WorkRally",
    "description": "P0 参考  ·  制作协同。项目、剧集、场次、资产、画布及批量生成，公开 CLI 文档可查，未登录实测。",
    "links": [
      "https://workrally.qq.com",
      "https://clawhub.ai/tencent-adm/skills/workrally"
    ],
    "status": "未下载",
    "note": "已查官网、官方CLI仓库及发布报道。本轮未取得可下载的独立平台演示；官网主脚本超过5MiB采集上限，未继续展开。",
    "leads": [
      {
        "title": "腾讯官方WorkRally仓库",
        "url": "https://github.com/Tencent/workrally"
      }
    ],
    "videos": []
  },
  {
    "id": "035",
    "title": "AutoRemesher",
    "description": "P1  ·  网格处理。自动四边形重拓扑与目标面数；不保证动画布线、UV 和贴图自动保留。",
    "links": [
      "https://github.com/huxingyi/autoremesher"
    ],
    "status": "未下载",
    "note": "原仓库附件与作者发布推文确认是静态图片，不是视频。本轮未找到符合要求的独立短演示。",
    "leads": [
      {
        "title": "作者发布推文，静态截图",
        "url": "https://x.com/jeremyhu2016/status/1294090448953196544"
      }
    ],
    "videos": []
  },
  {
    "id": "036",
    "title": "LTX Product Ad Style",
    "description": "P2  ·  产品广告。产品图驱动广告风格；包装文字和 Logo 需审片，不保证精确保留。",
    "links": [
      "https://huggingface.co/SOLRICKS/ltx-2.3-product-ad-style"
    ],
    "status": "已下载",
    "note": "作者产品广告风格样片，仅供观察效果；不据此承诺包装文字或商标精确保留。",
    "leads": [],
    "videos": [
      {
        "id": "036",
        "label": "产品广告风格样片",
        "url": "https://huggingface.co/SOLRICKS/ltx-2.3-product-ad-style/resolve/main/LTX-2.3_00013.mp4",
        "source": "https://huggingface.co/SOLRICKS/ltx-2.3-product-ad-style",
        "kind": "作者效果样片",
        "source_size": 16571444,
        "duration": 7.041667,
        "width": 2560,
        "height": 1408,
        "has_audio": true,
        "status": "downloaded",
        "file": "videos/036_产品广告风格样片.mp4",
        "bytes": 16571444,
        "sha256": "75e234ed31f878e72b223c9d65a3c5855e427b236deea4e359f7dfb78a30a9b4"
      }
    ]
  },
  {
    "id": "037",
    "title": "TRELLIS 与 Pixal3D GGUF 工作流",
    "description": "P1  ·  多视图建模。作者分享 Qwen 四视图与 3D 路线，JSON 未读取；量化与网格面数分别配置。",
    "links": [
      "https://www.reddit.com/r/comfyui/comments/1umsmau/gguf_workflows_for_trellis_2_and_pixel_3d/"
    ],
    "status": "未下载",
    "note": "原Reddit帖主要分享工作流，并链接两支YouTube教程。未取得对应短演示；第一支链接从6分33秒处开始，显然不是短片。001可作同类3D路线参考，但不计为本条已覆盖。",
    "leads": [
      {
        "title": "原帖链接的TRELLIS GGUF教程，长片不下载",
        "url": "https://www.youtube.com/watch?v=FuFm8zBHDWI&t=393s"
      },
      {
        "title": "原帖链接的Pixal3D GGUF教程",
        "url": "https://www.youtube.com/watch?v=LMmuhIwaeB4"
      }
    ],
    "videos": []
  },
  {
    "id": "038",
    "title": "Krea2 Realism V2",
    "description": "P2  ·  写实美术。纹理、光照及表情风格参考；与角色 LoRA 兼容性须实测，不用于时序一致性。",
    "links": [
      "https://huggingface.co/RudySen/Krea2-realism-V2"
    ],
    "status": "未下载",
    "note": "原模型是图像写实风格LoRA，页面以静态样图为主，本轮未找到独立短视频演示。",
    "leads": [],
    "videos": []
  },
  {
    "id": "039",
    "title": "ComfyUI OCIO",
    "description": "P2 扩展  ·  色彩与母版。OCIO 转换、ACES、LUT 和高精度输出。基础色彩元数据应在 P0 保存，专业输出后置。",
    "links": [
      "https://github.com/SlavaSexton/ComfyUI-OCIO"
    ],
    "status": "未下载",
    "note": "已查原仓库及作者LinkedIn公开介绍，展示内容为静态节点与对比截图，未找到对应独立视频。",
    "leads": [
      {
        "title": "作者OCIO说明与截图",
        "url": "https://www.linkedin.com/posts/slavasexton_comfyui-opencolorio-aces-activity-7479342225802436609-qrdP"
      }
    ],
    "videos": []
  },
  {
    "id": "040",
    "title": "LTX 10Eros 提取 LoRA",
    "description": "待核验  ·  模型组合。可确认是完整模型提取版并提到 GGUF 测试；独有能力未明确，暂不列创作需求。",
    "links": [
      "https://huggingface.co/adrepale/LTX2.3-10Eros-LoRA"
    ],
    "status": "未下载",
    "note": "原提取版LoRA未发现独立视频样例；独有能力本身仍待核验，不用基础模型或其他版本样片代替。",
    "leads": [],
    "videos": []
  },
  {
    "id": "041",
    "title": "Sun Direction Flux2 Klein",
    "description": "P1  ·  图片重打光。外景图加光照球控制太阳方向；强方向光先转阴天，室内及多灯能力未确认。",
    "links": [
      "https://huggingface.co/eric-venti-seeds/Sun-Direction-Lora-Flux2Klein9B"
    ],
    "status": "已下载",
    "note": "作者原始太阳方向对比GIF转MP4；该条是图片重打光，动画不表示模型直接生成视频。",
    "leads": [],
    "videos": [
      {
        "id": "041",
        "label": "太阳方向控制动画对比",
        "url": "https://huggingface.co/eric-venti-seeds/Sun-Direction-Lora-Flux2Klein9B/resolve/main/assets/newschwanstein.gif",
        "source": "https://huggingface.co/eric-venti-seeds/Sun-Direction-Lora-Flux2Klein9B",
        "kind": "作者动画对比",
        "source_size": 18086952,
        "conversion": "作者原始动画GIF转MP4；未新增画面；无音轨",
        "duration": 7.866667,
        "width": 844,
        "height": 472,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/041_太阳方向控制动画对比.mp4",
        "bytes": 971691,
        "sha256": "29a4a7cb6072d7f58035843660915f593442e8b9336bed516594c6f58991658f"
      }
    ]
  },
  {
    "id": "042",
    "title": "Film Space",
    "description": "P0 参考  ·  手机虚拟摄影。ARKit 实走运镜、人物占位与录像。移动端能力接入列 P1；轨迹导出未确认。",
    "links": [
      "https://github.com/maxprokopp/film-space"
    ],
    "status": "已下载",
    "note": "保留作者界面操作与相机演示两段原始动画，转为MP4便于本地播放。",
    "leads": [],
    "videos": [
      {
        "id": "042",
        "label": "FilmSpace手机虚拟片场演示",
        "url": "https://raw.githubusercontent.com/maxprokopp/film-space/HEAD/assets/film-space-demo.gif",
        "source": "https://github.com/maxprokopp/film-space",
        "kind": "作者动画演示",
        "source_size": 10758890,
        "conversion": "作者原始动画GIF转MP4；未新增画面；无音轨",
        "duration": 8.4,
        "width": 400,
        "height": 400,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/042_FilmSpace手机虚拟片场演示.mp4",
        "bytes": 564840,
        "sha256": "839837c92f3e22caeff66233874276fa4e54dfa542d896f5fd28fe9513e775f7"
      },
      {
        "id": "042",
        "label": "FilmSpace运镜参考演示",
        "url": "https://raw.githubusercontent.com/maxprokopp/film-space/HEAD/assets/film-space-camera-1.gif",
        "source": "https://github.com/maxprokopp/film-space",
        "kind": "作者动画演示",
        "source_size": 21800047,
        "conversion": "作者原始动画GIF转MP4；未新增画面；无音轨",
        "duration": 22.466667,
        "width": 400,
        "height": 400,
        "has_audio": false,
        "status": "downloaded",
        "file": "videos/042_FilmSpace运镜参考演示.mp4",
        "bytes": 1145532,
        "sha256": "54372ad79c963a58db02b832a6d8f72ca96ad50227b0935fb90bcd5a1604c0a3"
      }
    ]
  }
]