{ "total": 22, "cases": [ { "index": 1, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/5e774a6c50fe.jpg", "title": "ComfyUI-76-人像摄影极致写真流，超4K，数据解析！欢迎尝试！！！（小白福音）", "author": "", "body": "RunningHub工作流地址：https://www.runninghub.cn/post/2047526849235914754/?inviteCode=rh-v1234\n粉丝福利，注册填写rh-v1234，送1000点，每日登录100点，畅玩comfyui，超性能体验！\n本期视频模型+工作流：https://pan.quark.cn/s/ccae017425a4\nComfyui资料分享：https://pan.quark.cn/s/ba992dddc7c2\n调色节点详细介绍：https://www.\n\n[research_note] B站搜索结果；接口未返回有效发布时间，按教程内容作为专业工作流候选。", "images": [ "https://res.cybertogether.net/crawler/image/5e774a6c50fe.jpg" ], "url": "https://www.bilibili.com/video/BV12Ro5BcE9m", "note": "platform=bili | likes=0 | comments=0", "_raw": { "case_id": "bili_BV12Ro5BcE9m", "platform": "bili", "channel_content_id": "BV12Ro5BcE9m" }, "workflow": null, "capabilities": null }, { "index": 2, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/c85ff47a741f.jpg", "title": "Comfyui一键人像摄影工作流部署安装教程，从0到1教会你使用comfyui实现高级人像摄影", "author": "", "body": "comfyui一键人像写真工作流部署教程来了！网盘中包含需要用到的所有文件，你可以一键下载整合包运行，也可以跟着UP把遇到的问题全部解决，开始动手吧同学~\n\n其他工作流推荐：\n一键人像写真工作流【BV17F4m157Vg】\n一键换脸工作流【BV1qi421Z7V1】\n一键换脸工作流V2升级版【BV1Ni42197Fu】\n一键双人写真合照【BV13W421A7sy】\n\ncubiq作者instantid插件GitHub地址：https://github.com/cubiq/ComfyUI_InstantID\n\n[research_note] B站搜索结果；接口未返回有效发布时间，按教程内容作为专业工作流候选。", "images": [ "https://res.cybertogether.net/crawler/image/c85ff47a741f.jpg" ], "url": "https://www.bilibili.com/video/BV1rZ42187HK", "note": "platform=bili | likes=0 | comments=0", "_raw": { "case_id": "bili_BV1rZ42187HK", "platform": "bili", "channel_content_id": "BV1rZ42187HK" }, "workflow": null, "capabilities": null }, { "index": 3, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/a7d34b7183ae.jpg", "title": "[ComfyUI]粉丝想要“AI写真”？安排！Qwen+Nano 双方案，脸部完美复刻。", "author": "", "body": "根据参考图实现一键出写真集，分享开源闭源两套方案：\n\n工作流：https://www.runninghub.cn/post/1996923193854201857/?inviteCode=kol01-rh024\n网盘链接：https://pan.quark.cn/s/eb2b36813bbb\n\n粉丝福利，注册送1000点，每日登录100点，畅玩4090！体验48G的超能魅力！\n\n知识星球：https://t.zsxq.com/17kb0847s\n\n海外Plus会员：\nhttps://buy.52aibot\n\n[research_note] B站搜索结果；接口未返回有效发布时间，按教程内容作为专业工作流候选。", "images": [ "https://res.cybertogether.net/crawler/image/a7d34b7183ae.jpg" ], "url": "https://www.bilibili.com/video/BV1e42JBbE1u", "note": "platform=bili | likes=0 | comments=0", "_raw": { "case_id": "bili_BV1e42JBbE1u", "platform": "bili", "channel_content_id": "BV1e42JBbE1u" }, "workflow": null, "capabilities": null }, { "index": 4, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/6ebd4235f1f3.jpg", "title": "comfyui换背景人像融合写真Flux工作流 面部细节还原 婚纱摄影 大场景自由构图 超真实融合", "author": "", "body": "真实细节融合 透视关系明确 出图率超高的婚纱摄影人像融合换背景工作流\n\n工作流获取方式：https://igdy63tuaj.feishu.cn/docx/SEtwduKlNoxcZxxFj0AckfmPnsc?from=from_copylink\n\n[research_note] B站搜索结果；接口未返回有效发布时间，按教程内容作为专业工作流候选。", "images": [ "https://res.cybertogether.net/crawler/image/6ebd4235f1f3.jpg" ], "url": "https://www.bilibili.com/video/BV1nmqfYiER6", "note": "platform=bili | likes=0 | comments=0", "_raw": { "case_id": "bili_BV1nmqfYiER6", "platform": "bili", "channel_content_id": "BV1nmqfYiER6" }, "workflow": null, "capabilities": null }, { "index": 5, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/18231da63a24.jpg", "title": "Z-image+Klein 批量人像写真 新春特供 批量生成 制作属于你的专属写真 祝愿大家除夕快乐！年年岁岁好运常伴，财旺势旺身体棒！", "author": "", "body": "工作流：「鳌拜」Z-image+klein批量写真工作流 新春特供\n体验地址：https://www.runninghub.ai/post/2023430928676163585/?inviteCode=rh-v1516\n通过我的链接🔗注册可以领取1000RH币游玩~\n\n工作流本地部署（夸克）：https://pan.quark.cn/s/238db95eda63\n（🔔：配合视频教程食用更佳哦~）\n\nhttps://huggingface.co/wikeeyang/Flux2-Klein-9B-True-\n\n[research_note] B站搜索结果；接口未返回有效发布时间，按教程内容作为专业工作流候选。", "images": [ "https://res.cybertogether.net/crawler/image/18231da63a24.jpg" ], "url": "https://www.bilibili.com/video/BV1krZMBzEn9", "note": "platform=bili | likes=0 | comments=0", "_raw": { "case_id": "bili_BV1krZMBzEn9", "platform": "bili", "channel_content_id": "BV1krZMBzEn9" }, "workflow": null, "capabilities": null }, { "index": 6, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/ef2f4ace3a0f.jpg", "title": "拒绝AI感，Comfyui超细腻人物皮肤质感工作流", "author": "", "body": "100+课时《全面掌握Comfyui》系统课程 https://www.bilibili.com/cheese/play/ss61694\n\n本期教程用到的模型、素材、工作流：\n度：https://pan.baidu.com/s/1CGvtFsZnizJPnFNsml1bIw?pwd=cgmi\n夸：https://pan.quark.cn/s/1fe39155ccff\n\n本期工作流Runninghub直接在线运行：https://www.runninghub.cn/post/1933833938685054\n\n[research_note] B站搜索结果；接口未返回有效发布时间，按教程内容作为专业工作流候选。", "images": [ "https://res.cybertogether.net/crawler/image/ef2f4ace3a0f.jpg" ], "url": "https://www.bilibili.com/video/BV1XSMezxEdU", "note": "platform=bili | likes=0 | comments=0", "_raw": { "case_id": "bili_BV1XSMezxEdU", "platform": "bili", "channel_content_id": "BV1XSMezxEdU" }, "workflow": null, "capabilities": null }, { "index": 7, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/3a25ccc033b5.jpg", "title": "ComfyUI AI Influencer Tutorial Part 1: How to Get a Lifelike Look with Z-Image Turbo", "author": "AI Influencer Creator", "body": "Learn how to create a lifelike AI influencer using Z-Image Turbo inside ComfyUI.\nThis tutorial focuses on achieving true photorealism through natural lighting, subtle imperfections, and authentic photo aesthetics — the key to removing the typical AI look.\nAI Influencer Tutorial Series — Part 1\n\n🔹 ComfyUI Workflow\nhttps://github.com/influencerbyai/com...\n\n🔹 Z-Image Turbo (GGUF)\nhttps://huggingface.co/unsloth/Z-Imag...\n🔹 vae\nhttps://huggingface.co/Comfy-Org/z_im...\n\n💻 No GPU? No Problem\nYou can try Z-Image Turbo online for free:\n \n\nKey Discovery\nThe biggest realism boost came from using styles like:\n Phone Photo \n Selfie \n Casual Photo \n Portra Film Photo \nAdding raw-photo language to prompts made a massive difference: raw texture, natural imperfections, shallow depth of field,\nslightly imperfect framing, unedited look.\n\nThis removes the \"AI plastic skin\" effect almost immediately.\n\nTest Prompts (Try Them Yourself)\nUse the prompts below with the workflow and observe how small changes create completely different influencer styles:\n\nPrompt 1: Photorealistic candid photo of a 22-year-old young Caucasian woman with fair skin, natural light freckles across her nose and cheeks, bright blue eyes, long wavy ash-blonde hair tied in a high messy ponytail, doing glute bridge in modern gym, wearing tight black sports bra and high-waist leggings, confident playful smile looking at camera, sweat glow on skin, bright gym lighting with natural window light, energetic and fresh atmosphere, vertical portrait, raw mobile photo aesthetic, natural imperfections, gym mirror reflection slightly visible, highly realistic, sharp focus on face and body, raw texture, natural imperfections, shallow depth of field, sharp focus on subject, slightly imperfect framing, raw photo, unedited look\n\nPrompt 2:a young Caucasian woman is standing on a rocky beach at night. She is wearing a white bikini and has her hair pulled up in a ponytail. Her left hand is resting on her head, while her right hand is placed on her hip. She is smiling and looking over her shoulder at the camera. The background features a dark sky with some clouds and the ocean, which is illuminated by green lights. The woman appears to be enjoying her time at the beach, raw texture, natural imperfections, shallow depth of field, sharp focus on subject, slightly imperfect framing, raw photo, unedited look\n\n Prompt 3: Influencer style, joyful beach selfie of a vibrant Brazilian beauty with glowing golden-tan skin and voluminous curly dark hair, standing barefoot on Ipanema Beach at golden hour. She wears a colorful bikini top and flowy white linen skirt, laughing naturally while wind blows her hair. Turquoise ocean and Sugarloaf Mountain in background. Raw texture, natural imperfections, shallow depth of field, sharp focus on subject, slightly imperfect framing, raw photo, unedited look, warm sun-kissed skin glow and salty breeze energy.\n\n Prompt 4: A 20-year-old confident American beauty with lightly sun-kissed skin, bright blue eyes, and long wavy honey-blonde hair in a messy high ponytail with loose strands, stands in front of a gym mirror taking a casual post-workout selfie. Fresh, energetic yet relaxed morning vibe. She wears a fitted light grey sports bra and high-waisted black seamless biker shorts that sculpt her toned body, visible abs and athletic legs. Pose: one hand holding iPhone, the other on her hip, slight body tilt, confident but soft smile, natural flushed skin. Modern gym interior with soft natural window light. Raw texture, natural imperfections, shallow depth of field, sharp focus on subject, slightly imperfect framing, raw photo, unedited look, fresh and authentic fitness influencer energy.\n\n Prompt 5: Cinematic ultra-realistic lifestyle shot of a vibrant Brazilian beauty, walking through a private airport lounge, looking at camera, minimalist chic travel outfit in pastel tones, designer bag casually over shoulder, glancing at phone, sunlight streaming through windows, natural shadows, candid movement, relaxed and confident energy, subtle hair highlights reflecting light, realistic skin texture, raw texture, natural imperfections, shallow depth of field, sharp focus on subject, slightly imperfect framing, raw photo, unedited look\n\nIf this tutorial helped you, consider liking and subscribing — it really supports the channel.\nLet me know in the comments which result looked the most realistic to you 👇\n#comfyui #zimage #aiimagegenerator #aimodel #aiworkflow #stablediffusion #freeaitools #freeaitools #aitutorial #opensourceai #lora #aiart #aigenerated #generativeai #texttoimage #photorealism #aicreator #aigenerated", "images": [ "https://res.cybertogether.net/crawler/image/3a25ccc033b5.jpg" ], "url": "https://www.youtube.com/watch?v=1mGQgEnaLfA", "note": "platform=youtube | likes=42 | comments=3", "_raw": { "case_id": "youtube_1mGQgEnaLfA", "platform": "youtube", "channel_content_id": "1mGQgEnaLfA" }, "workflow": null, "capabilities": null }, { "index": 8, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/bc68c87ac40f.jpg", "title": "Turn Waxy AI Portraits into Photorealistic Masters in Seconds (Custom Workflow)", "author": "Prompt Mastery", "body": "Description:\n\n👇 GET THE WORKFLOW HERE:\nhttps://drive.google.com/file/d/1oewD...\n(This is the Step-Injection workflow featured in the video. No plugins required.)\n\n🔥 IN THIS VIDEO:\nAre your AI portraits looking like waxy plastic dolls? Stop upscaling pixels. The secret to hyper-realistic skin texture isn't in the resolution—it's in the Latent Steps.\n\nToday, I reveal my exclusive \"Step Control\" method in ComfyUI by using Zimage turbo latent repaint function. I'll show you how to:\n\nInject real human skin texture into ANY Stable Diffusion image.\n\nUse \"Start/End Step\" controls to surgically fix details.\n\nTurn a \"plastic\" generation into a photorealistic masterpiece in seconds.\n\nNo expensive plugins. No 20-step Denoise nightmare. Just pure control.\n\n📺 WATCH NEXT (My Local AI Studio Series):\n\nIf you are new to running these models locally, check out my previous guides to get set up:\n\nTurn Blurry Images Into 4K Masterpieces With This FLUX Klein Workflow\nMy complete breakdown on how to use Flux Klein + CVR2 to restore details from low-res messy images.\n🔗   • Turn Blurry Images Into 4K Masterpieces Wi...  \n\nFlux 2 Klein + Dark Beast (The Ultimate Photorealism):\nMy latest breakdown of the Dark Beast LoRA for insane realism.\n🔗    • Flux 2 Klein + Dark Beast: The Ultimate Ph...  \n\nFlux 2 Klein \"Secret\" Edit Hack:\nThe censorship hack and editing trick that got everyone talking. (Must watch if you want full control).\n🔗    • Flux 2 Klein Censorship Hack #photoshop #i...  \n\n🛠️ TOOLS & RESOURCES:\n\nComfyUI: https://github.com/comfyanonymous/Com...\n\nModel Used: [Checkpoints like Juggernaut XL or RealVisXL recommended]\n\nLoRA Used (Crucial for Style):\n\nDarkbeast Zimage LoRA: https://civitai.com/api/download/mode...\n\n⚠️ IMPORTANT: Set LoRA strength to 0.3 for the best results (as shown in video).\n\n⏱️ TIMESTAMPS:\n\n\n👋 CONNECT WITH ME:\n\nSubscribe for more AI Workflows:    / @promptmastery  \n#comfyui #ai workflow\n\n00:00 - Result Showcase: Turning \"Waxy\" Portraits into Reality\n01:38 - The Problem: Fixing Flux Client's \"Snake Skin\" Texture\n05:06 - ⚠️ CRITICAL LOGIC: How \"Start Step\" & \"End Step\" Work\n07:19 - Live Demo: Step 5 vs. Step 3 (Makeup vs. Full Redraw)\n10:36 - Advanced Trick: Multi-Pass Generation (Preserve 100% Face)\n12:51 - Second Pass Demo: Adding \"Microscope Level\" Details\n15:06 - LoRA Settings (Dark BZ) & Final Comparison\n16:08 - Conclusion & Next Video Preview\n\nThis video shows the creation of a stunning 200-megapixel picture, highlighting its incredible image resolution and detail. We zoom in to demonstrate the exceptional quality, which is comparable to 4k video ultra hd, even revealing intricate skin texture. This process showcases the power of an advanced ai image upscaler and repaint, pushing the boundaries of what's possible in digital imaging.", "images": [ "https://res.cybertogether.net/crawler/image/bc68c87ac40f.jpg" ], "url": "https://www.youtube.com/watch?v=yA0rEz1ByCg", "note": "platform=youtube | likes=284 | comments=74", "_raw": { "case_id": "youtube_yA0rEz1ByCg", "platform": "youtube", "channel_content_id": "yA0rEz1ByCg" }, "workflow": null, "capabilities": null }, { "index": 9, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/1c07c8b5e471.jpg", "title": "Advanced Face Swap Workflow for Consistent Portrait Editing", "author": "congXS", "body": "#comfyui #comfyuitutorial #ai #faceswap #runninghub #workflow", "images": [ "https://res.cybertogether.net/crawler/image/1c07c8b5e471.jpg" ], "url": "https://www.youtube.com/watch?v=0I51nQq11e4", "note": "platform=youtube | likes=7 | comments=3", "_raw": { "case_id": "youtube_0I51nQq11e4", "platform": "youtube", "channel_content_id": "0I51nQq11e4" }, "workflow": null, "capabilities": null }, { "index": 10, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/28bebd686f4f.jpg", "title": "Photoshop Is DEAD - AI Photo Restoration That Actually Works (ComfyUI Flux Kontext)", "author": "Aiconomist", "body": "🚀 Master AI Creation (2026 AI Influencer Course):   / 2048552  \n------------------------------------------------------------------\nDownload Links:\n⬇️ Workflow and All Resources: https://ko-fi.com/s/b697509674\n------------------------------------------------------------------\nWebsite: https://pixelailabs.com\nFor Business: ➜ aiconomist@pixelailabs.com\n------------------------------------------------------------------\n💻 Hardware Requirements:\nNVIDIA GPU (8GB+ VRAM recommended)\n16GB+ System RAM\nComfyUI Installation\n🔧 Compatible with:\nRTX 4060 | RTX 3060 | RTX 4090 | RTX 5090 | RTX 4080 | RTX 3090 | A6000 | H100\n------------------------------------------------------------------\n\n⏰ Timestamps:\n0:00 - Introduction \n0:43 - Flux Kontext Info & Setup\n1:30 - Example 1\n5:28 - Example 2\n6:08 - Example 3\n------------------------------------------------------------------\nLearn how to restore damaged photos and vintage images using AI with this comprehensive Flux Kontext ComfyUI tutorial. This free photo restoration workflow allows you to repair old photographs, restore vintage family photos, and fix damaged images in seconds without expensive software like Photoshop. Whether you need to restore black and white photos, repair torn pictures, colorize old images, or enhance vintage portraits, this AI image restoration technique delivers professional results instantly. The Flux Kontext model combined with ComfyUI provides powerful photo repair capabilities for restoring faded photographs, fixing scratched images, and bringing old family memories back to life. Download the complete image restoration workflow for free and transform your damaged photos into stunning restored masterpieces using this cutting-edge AI photo enhancement technology. Perfect for anyone looking to restore family photographs, repair vintage images, or learn advanced photo restoration techniques using artificial intelligence and ComfyUI workflows.\n\n#WanStepDistill #AIVideo #ComfyUI #VideoGeneration #LoRA #MachineLearning #AITools #VideoAI #StableDiffusion #CreativeAI #TechTutorial #AIWorkflow", "images": [ "https://res.cybertogether.net/crawler/image/28bebd686f4f.jpg" ], "url": "https://www.youtube.com/watch?v=2zng0kwIjbw", "note": "platform=youtube | likes=757 | comments=57", "_raw": { "case_id": "youtube_2zng0kwIjbw", "platform": "youtube", "channel_content_id": "2zng0kwIjbw" }, "workflow": null, "capabilities": null }, { "index": 11, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/7cc2be16eeff.jpg", "title": "BEST AI Skin Fix and Upscaling Workflow | Full Body Skin Refiner ComfyUI Workflow", "author": "Silent Snow", "body": "This is, by far, the best full body AI skin fixing and upscaling workflow we have ever created. All local, low VRAM, Enhancor AI ComfyUI workflow alternative that delivers as promised.\n\nBuilding off of our original Mega Skin Refiner workflow, we heard the opinions and comments. That's why with this new set of workflows, we've aimed to ensure that you are not only getting better results than your input image every time guaranteed, but that this workflow is as light weight, easy to use, and supportive as possible.\n\nThis time, we're providing manual masking options right out of the box, whether you want to make your mask using the built in ComfyUI mask editor, or upload your own from your files, we've got you covered.\n\nAnd for those of you aiming to run this workflow locally, we've also created a lighter weight variation that should work on most PC's with a GPU and a pulse.\n\nWe guarantee that no matter which workflow you choose, you will receive at least one image that is better than the original image you put in.\n\nCheck out the workflows below!\n\nLinks 🔗\n\nPaid Workflows: https://www.patreon.com/posts/new-ope...\n\nFree Workflow: https://www.patreon.com/posts/new-ope...\n\nPrevious Skin Fix Tutorial:    • AI Skin Fixing and Upscaling | Full Body S...  \n\n#comfyui #skinrefiner #aiplasticskin #plasticskinfix #enhancorai\n#enhancor", "images": [ "https://res.cybertogether.net/crawler/image/7cc2be16eeff.jpg" ], "url": "https://www.youtube.com/watch?v=fQadB_s-UV0", "note": "platform=youtube | likes=336 | comments=41", "_raw": { "case_id": "youtube_fQadB_s-UV0", "platform": "youtube", "channel_content_id": "fQadB_s-UV0" }, "workflow": null, "capabilities": null }, { "index": 12, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/4e5d529b49dd.jpg", "title": "Z-Image Turbo vs FLUX.2 Klein 9B Photorealism Face-Off | ComfyUI AI Influencer Tutorial Part 2", "author": "AI Influencer Creator", "body": "In Part 2 of the AI Influencer series, we directly compare Z-Image Turbo and FLUX.2 Klein 9B inside ComfyUI to see which model delivers stronger photorealism.\n🔹 ComfyUI Workflow\nFLUX.2 Klein 9B:\nhttps://github.com/influencerbyai/com...\nZ-Image Turbo:\nhttps://github.com/influencerbyai/com...\n🔹 Flux 2 Klein 9B Model:\n https://huggingface.co/black-forest-l...\n🔹 VAE:\n https://huggingface.co/Comfy-Org/vae-...\n💻 No GPU? No Problem\nYou can try Z-Image Turbo online for free\nTest Prompts (Try Them Yourself)\nYou can use the prompts below with the FLUX.2 Klein 9B workflow to reproduce the results shown in this video.\nPrompt 1: \nYOUR CONTEXT:\nThe subject takes a selfie with their limb outstretched. The selfie has android phone cam-quality.\nThe selfie features a slightly tilted, close-up composition, sharp and complex backgrounds, and a spontaneous moment that feels immediate and authentic. Subtle film grain and a soft halo add a touch of authenticity and nostalgia. Skin tones are realistic and warm, reflecting natural light.\nTHE SELFIE:\nPhotorealistic candid photo of a 22-year-old young Caucasian woman with fair skin, natural light freckles across her nose and cheeks, bright blue eyes, long wavy ash-blonde hair tied in a high messy ponytail, doing glute bridge in modern gym, wearing tight black sports bra and high-waist leggings, confident playful smile looking at camera, sweat glow on skin, bright gym lighting with natural window light, energetic and fresh atmosphere, vertical portrait, raw mobile photo aesthetic, natural imperfections, gym mirror reflection slightly visible, highly realistic, sharp focus on face and body, raw texture, natural imperfections, shallow depth of field, sharp focus on subject, slightly imperfect framing, raw photo, unedited look\nPrompt 2:\nYOUR CONTEXT:\nYou are a photographer who appreciates the classic aesthetic of Kodak Portra film.\nYour photograph emulates that look, known for its soft colors, fine grain, and natural skin tones. The image features a subtle warmth, with a focus on accurate color rendition and a gentle, diffused glow. Highlights are smooth and creamy, while shadows retain detail. Minimal post-processing is applied to preserve the organic, film-like quality.\nYOUR PHOTOGRAPH:\na young Caucasian woman is standing on a rocky beach at night. She is wearing a white bikini and has her hair pulled up in a ponytail. Her left hand is resting on her head, while her right hand is placed on her hip. She is smiling and looking over her shoulder at the camera. The background features a dark sky with some clouds and the ocean, which is illuminated by green lights. The woman appears to be enjoying her time at the beach, raw texture, natural imperfections, shallow depth of field, sharp focus on subject, slightly imperfect framing, raw photo, unedited look\nPrompt 3: \nYOUR CONTEXT:\nThe subject takes a selfie with their limb outstretched. The selfie has android phone cam-quality.\nThe selfie features a slightly tilted, close-up composition, sharp and complex backgrounds, and a spontaneous moment that feels immediate and authentic. Subtle film grain and a soft halo add a touch of authenticity and nostalgia. Skin tones are realistic and warm, reflecting natural light.\nTHE SELFIE:\nInfluencer style, joyful beach selfie of a vibrant Brazilian beauty with glowing golden-tan skin and voluminous curly dark hair, standing barefoot on Ipanema Beach at golden hour. She wears a colorful bikini top and flowy white linen skirt, laughing naturally while wind blows her hair. Turquoise ocean and Sugarloaf Mountain in background. Raw texture, natural imperfections, shallow depth of field, sharp focus on subject, slightly imperfect framing, raw photo, unedited look, warm sun-kissed skin glow and salty breeze energy.\n\n#comfyui #zimage #zimageturbo #aimodel #aiworkflow #stablediffusion #aitutorial #opensourceai #aiart #aigenerated #generativeai #texttoimage \n#photorealism #aiimagegenerator #aiinfluencer #aiinfluencertutorial\n#aiphotography #realisticai #aiimages #aicreator #comfyuiworkflow #aicomparison #airealism #aicontentcreation #aitools", "images": [ "https://res.cybertogether.net/crawler/image/4e5d529b49dd.jpg" ], "url": "https://www.youtube.com/watch?v=MJ4sXyOOGbA", "note": "platform=youtube | likes=32 | comments=5", "_raw": { "case_id": "youtube_MJ4sXyOOGbA", "platform": "youtube", "channel_content_id": "MJ4sXyOOGbA" }, "workflow": null, "capabilities": null }, { "index": 13, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/a40abd6e2aa6.jpg", "title": "最强人物一致性工作流 comfyui教程", "author": "", "body": "最强人物一致性工作流 comfyui教程#comfyui教程[话题]# #comfyui[话题]# #Lib布道师计划[话题]# #LiblibAI[话题]# #lora[话题]# #小进步学年[话题]#", "images": [ "https://res.cybertogether.net/crawler/image/a40abd6e2aa6.jpg" ], "url": "https://www.xiaohongshu.com/explore/6784e5d80000000020018880", "note": "platform=xhs | likes=1453 | comments=None", "_raw": { "case_id": "xhs_3", "platform": "xhs", "channel_content_id": "6784e5d80000000020018880" }, "workflow": null, "capabilities": null }, { "index": 14, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/2b5bd52d2790.jpg", "title": "Comfyui 超级高清放大，人像增强Mix", "author": "", "body": "相比其他工作流，这个工作流的优势是将传统Ps中的细节增强流程，通过Comfyui实现，同时叠加Ai生成及细节增强的能力，有了很强的增强效果。\n\t\n对工作流有优化想法的，可以在评论区讨论\n#comfyui[话题]#  #comfyui教程[话题]#  #Ai[话题]#  #A神[话题]#  #stablediffusion[话题]#  #ai绘画[话题]#  #设计师[话题]#  #视觉设计[话题]#  #人像修图[话题]#", "images": [ "https://res.cybertogether.net/crawler/image/2b5bd52d2790.jpg", "https://res.cybertogether.net/crawler/image/c41e98ef3847.jpg", "https://res.cybertogether.net/crawler/image/ed1faaca98d9.jpg", "https://res.cybertogether.net/crawler/image/228d8fad2585.jpg", "https://res.cybertogether.net/crawler/image/71f9934c1e59.jpg", "https://res.cybertogether.net/crawler/image/a2b3f47499ef.jpg", "https://res.cybertogether.net/crawler/image/e8393b24addd.jpg", "https://res.cybertogether.net/crawler/image/a426c8cb545a.jpg", "https://res.cybertogether.net/crawler/image/4ee415edab77.jpg", "https://res.cybertogether.net/crawler/image/a61c098cbf59.jpg", "https://res.cybertogether.net/crawler/image/6da5fd1e1794.jpg" ], "url": "https://www.xiaohongshu.com/explore/66688293000000000d00f2aa", "note": "platform=xhs | likes=94 | comments=None", "_raw": { "case_id": "xhs_6", "platform": "xhs", "channel_content_id": "66688293000000000d00f2aa" }, "workflow": null, "capabilities": null }, { "index": 15, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/5f35825771c5.jpg", "title": "超变态的AI一键写真！免费分享工作流教程！", "author": "", "body": "超变态的AI一键写真！免费分享工作流教程！\n#AI绘画[话题]# #ai[话题]# #写真[话题]# #comfyui[话题]# #comfyui工作流[话题]# #comfyui教程[话题]# #AI绘画教程[话题]# #comfyui产品一键换背景[话题]# #comfyui写真工作流[话题]#", "images": [ "https://res.cybertogether.net/crawler/image/5f35825771c5.jpg" ], "url": "https://www.xiaohongshu.com/explore/67ce5201000000000302ae7c", "note": "platform=xhs | likes=16 | comments=None", "_raw": { "case_id": "xhs_7", "platform": "xhs", "channel_content_id": "67ce5201000000000302ae7c" }, "workflow": null, "capabilities": null }, { "index": 16, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "https://res.cybertogether.net/crawler/image/15ffeba62ad1.jpg", "title": "新人必学的ComfyUI AI写真人物面部处理技巧", "author": "", "body": "如果你和我一样正在学习  #ComfyUI[话题]# ，今天的分享应该会对你有帮助（以AI写真职业形象照为示例）\n-  人物一致性选择 InstantID 还是 PulID ？\n-- InstantID支持参考原图的 KPS(Face KeyPoints 理解为脸部的姿势控制)或指定KPS，且质感略微比 PulID 强，如果你需要参考原图的人脸构图出图，那么选他就对了\n-- PulID 不需要也不支持KPS，因为它只影响结果的面部，而且效果更平滑，符合我的审美 如果你需要随机的面部姿势构图，或更想通过其他手段控制构图/姿势，我大力推荐你使用 PulID 如果你需要生成基于原图的性别，PulID 还需要提示词辅助，不过 ComfyUI 很容易搭出性别检测\n·\n-  如何在 #comfyui[话题]# 实现性别检测\n-- 最简单的就是 #WD14-Tagger，阈值设高，第 1-2 个提示词一定是 1girl/1boy，但是不够高效，3080 下大约需要 4-5s，当然如果你本身就在用 WD14-Tagger 那就无所谓了！我更推荐的是，ImpactHFTransformersClassifierProvider 并配合其他节点，如图2\n·\n-  一张图片上多个人物面孔，如何单独处理一张面孔？\n-- 尽管支持多ID，但是多个人脸会占用更多的资源、且不符合用户的预期，实现起来超级简单，只需一个 FaceBoundingBox 节点，记得将其 index 设置为 0 值得一提的是 padding 一定要把控好，太少缺少脑壳，太多容易把贴贴的两个人搞一起，.8 是个不错的选择 如图3\n·\n-  一个冷知识，人类对陌生人脸的识别不是很擅长，得益于 FaceEmbedDistance 节点，你可以很容易通过AI 来识别生成的像还是不像，可以在调试时提供参考价值（分数越小越好）！ 左边是 PulID 直出，中间是原图 右边是额外使用了 #IPAdapter Plus Face，具体操作如图4  #三花ai[话题]# #AI[话题]#  #好用的AI工具分享[话题]#  #学习[话题]#  #comfyui教程[话题]#  #comfyui工作流[话题]#  #三花AI[话题]#  #comfyui技巧[话题]#  #AI写真[话题]#  #职业照[话题]#", "images": [ "https://res.cybertogether.net/crawler/image/15ffeba62ad1.jpg", "https://res.cybertogether.net/crawler/image/66d0a0168a8f.jpg", "https://res.cybertogether.net/crawler/image/28c2fe32c855.jpg", "https://res.cybertogether.net/crawler/image/e9312a8ae8f1.jpg" ], "url": "https://www.xiaohongshu.com/explore/66700eae00000000060043b7", "note": "platform=xhs | likes=25 | comments=None", "_raw": { "case_id": "xhs_10", "platform": "xhs", "channel_content_id": "66700eae00000000060043b7" }, "workflow": null, "capabilities": null }, { "index": 17, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "", "title": "Photoshoot with GPT image 2.0", "author": "craftian_keskin", "body": "Photoshoot with GPT image 2.0\n\nPrompt:\n\n { \n \"scene_type\": \"smartphone fashion portrait series\",\n ...\n\n[research_note] X搜索结果，用作英文 prompt/工作流趋势样本；未作为专业教程主证据。", "images": [], "url": "", "note": "platform=x | likes=475 | comments=15", "_raw": { "case_id": "x_3", "platform": "x", "channel_content_id": "idx3" }, "workflow": null, "capabilities": null }, { "index": 18, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "", "title": "AI Detailed Prompt:", "author": "itsphotogptai", "body": "AI Detailed Prompt:\nUltra-realistic cinematic portrait of [REFERENCE PERSON] preserve facial structu...\n\n[research_note] X搜索结果，用作英文 prompt/工作流趋势样本；未作为专业教程主证据。", "images": [], "url": "", "note": "platform=x | likes=14 | comments=4", "_raw": { "case_id": "x_4", "platform": "x", "channel_content_id": "idx4" }, "workflow": null, "capabilities": null }, { "index": 19, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "", "title": "I just built a new ComfyUI workflow combining LTX 2.3 and Prompt Relay. It gives", "author": "aistudynowcom", "body": "I just built a new ComfyUI workflow combining LTX 2.3 and Prompt Relay. It gives us absolute control...\n\n[research_note] X搜索结果，用作英文 prompt/工作流趋势样本；未作为专业教程主证据。", "images": [], "url": "", "note": "platform=x | likes=11 | comments=1", "_raw": { "case_id": "x_9", "platform": "x", "channel_content_id": "idx9" }, "workflow": null, "capabilities": null }, { "index": 20, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "", "title": "{", "author": "sidona", "body": "{\n \"portrait_prompt\": {\n \"subject\": {\n \"description\": \"Based on , create an ...\n\n[research_note] X搜索结果，用作英文 prompt/工作流趋势样本；未作为专业教程主证据。", "images": [], "url": "", "note": "platform=x | likes=2 | comments=0", "_raw": { "case_id": "x_11", "platform": "x", "channel_content_id": "idx11" }, "workflow": null, "capabilities": null }, { "index": 21, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "", "title": "Entire fashion shoot replaced by one prompt.", "author": "Noor_ul_ain43", "body": "Entire fashion shoot replaced by one prompt.\nWelcome to the AI era. \n\nPrompt ⤵️ \n\nA high-fashion bla...\n\n[research_note] X搜索结果，用作英文 prompt/工作流趋势样本；未作为专业教程主证据。", "images": [], "url": "", "note": "platform=x | likes=129 | comments=19", "_raw": { "case_id": "x_14", "platform": "x", "channel_content_id": "idx14" }, "workflow": null, "capabilities": null }, { "index": 22, "category": "", "user_kept": false, "user_comment": "", "description": "", "method": "", "cover": "", "title": "公众号原文：Nano Banana Pro vs Z-Image-Turbo 使用教程和深度评测", "author": "公众号线索（网页反查）", "body": "2025-12-07 公众号原文线索。文章比较 Google Nano Banana Pro 与阿里 Z-Image-Turbo，并给出 Z-Image-Turbo 在 ComfyUI 中的部署路径：下载 text_encoder / diffusion_model / vae，导入 workflow JSON 运行；同时提到 GGUF 版本可降低内存要求。对 AI 人物写真组图的价值：Z-Image-Turbo 被描述为照片级真实感、皮肤质感和发丝细节较强，适合作为真实感人像静帧底模候选。\n\n[research_note] content-search 公众号接口本轮返回空；此条由网页搜索反查到 mp.weixin.qq.com 原文链接后补充。参考转载页：https://blog.csdn.net/djstavaV/article/details/155583509", "images": [], "url": "https://mp.weixin.qq.com/s?__biz=MzA5MTgzNzg4NQ==&mid=2455767882&idx=1&sn=a7bcb60f431a3cb1623bab39ac83fee8&chksm=863f2cead6f7d30ddbcc54f6d997575fff3516190a3ee5473df27e1f90255442736fee8528ef&scene=126&sessionid=0", "note": "platform=gzh | likes=0 | comments=0", "_raw": { "case_id": "gzh_web_zimage_turbo_nanobanana_20251207", "platform": "gzh", "channel_content_id": "zimage_turbo_nanobanana_20251207" }, "workflow": null, "capabilities": null } ] }