Skip to content
Date and source

Dates follow the news timeline in Beijing time and include both endpoints.

TodayNov 19Thu2 items
  1. Stable Diffusion Art 教程65

    MiniMax H3 Image Edit Workflow Tutorial: Converting Anime Images to Realistic Photos

    Stable Diffusion Art publishes a tutorial showing how to convert anime images into realistic photos in ComfyUI using the MiniMax H3 Image Edit workflow. The workflow uses only native nodes, preserving character pose and composition through the vision-language pathway and reference latent pathway, without needing a ControlNet stack. Optionally enable the Qwen 3.5 prompt enhancer, supporting scene-level custom edits.

    Why it matters: This workflow uses native nodes to convert anime to realistic style, eliminating the need for a ControlNet stack. It can be applied to most anime characters and scenes without requiring additional edit checkpoints.

  2. note · tidaa · 图像视频创作分享31

    Midjourney prompt generator ver.6.2 released, dual-person mode greatly evolved, and facial component settings added.

    Midjourney prompt generator ver.6.2 released, dual-person mode changed to Person 1/Person 2 tabs, allowing separate settings for character, hairstyle, clothing, etc. Added 8 facial component settings (contour, eyes, eyebrows, nose, lips, etc.) and a one-click swap button. Mobile version Mobile v6.2 updated simultaneously, tool runs completely locally (HTML/JavaScript).

Nov 18Wed
  1. Stable Diffusion Art 教程74

    MiniMax H3 Image generates native 16MP images in ComfyUI (with workflow)

    MiniMax H3 Image can natively generate approximately 16.8MP images in ComfyUI, completed in 6 steps, without external super-resolution. The tutorial provides a complete workflow JSON and gives quantized model options: diffusion model 21GB, text encoder 15.7GB, image VAE 1.17GB, and can run on a 24GB GPU.

    Why it matters: The tutorial provides a complete ComfyUI workflow and quantized model options, allowing local reproduction of native 16.8MP image generation, eliminating the external super-resolution step.

  2. The Verge · AI59

    Suno launches Visuals feature, which can generate videos and images for AI music.

    Suno launches Visuals feature, allowing users to turn songs into AI-generated video clips, automatically lip-sync lyrics, or generate images from prompts for use as thumbnails and album covers. Visuals supports web, iOS, and Android, with videos up to 15 seconds long, driven by the H3 Move AI model based on MiniMax H3, and includes a video editing timeline. Visuals uses separate Visuals credits, separate from song credits, requiring users to pay extra.

  3. 可灵官方博客65

    Kling releases Kling Image 4.0, focusing on realism and native 8K generation.

    Kling has released its next-generation image generation model, Kling Image 4.0, now available on Kling Creative Studio and open to enterprises and developers via API. The model is based on a native unified understanding and generation architecture, supporting native 8K image generation, text rendering of up to 2000 characters per image, generation of up to 10 consistent-style images from a single prompt, and long-image generation such as e-commerce detail pages. It also offers 110+ preset color grading styles, 70+ preset commercial fonts, and support for seven languages.

    Why it matters: The new model unifies visual understanding and image generation in the same autoregressive framework, and its native 8K and long-text rendering capabilities are worth creators evaluating as a replacement for existing image generation workflows.

  4. comfyorg · YouTube 创作视频20

    Model Jam:Seedream 5.5

    ComfyUI team livestreams model improvisation with Seedream 5.5, exploring its capabilities starting from a blank canvas. The event covers first impressions of Seedream 5.5, real-time prompt creation, building reusable ComfyUI workflows, and live experiments. Viewers can submit prompts and questions during the livestream.

  5. note · kongo_jun · 图像视频创作分享62

    Trying the 2step LoRA of MiniMax-H3-Image and P-H3-Image in ComfyUI

    The author configured and tested the 2step LoRA of MiniMax-H3-Image and P-H3-Image in ComfyUI; generation times were 16 seconds and 19 seconds respectively, with VRAM at 21GB for both. The author feels that for T2I, P-H3-Image is more to their taste, while for image editing, MiniMax-H3-Image is more to their taste, and notes that the speed-up LoRA seems useful in high-resolution generation.

Nov 17Tue
  1. note · bochibochi_ai · 图像视频创作分享41

    Google Pics updated to Nano Banana 2.1, reducing degradation of text within images and localized edits

    On November 16, Google announced that its image generation and editing tool Google Pics has been updated to the Nano Banana 2.1 model. The new model maintains the same processing speed while improving image quality and realism, and enhancing editing consistency and text rendering; during some edits, elements such as backgrounds are less prone to degradation. The update requires no user settings and is available to paid plans such as Google AI Pro/Ultra, rolling out gradually over 1–3 days starting November 16.

  2. note · ai_freak · 图像视频创作分享58

    Seedream 5.5 update experience sharing: four improvements including realistic texture and local editing

    Seedream 5.5 update brings four improvements: more natural realistic texture, more stable local editing, clearer small text, and transparent background generation. The author tested it on Higgsfield and found that local editing can modify only the specified area without affecting other parts, and transparent backgrounds can directly generate LINE stickers and logos.

  3. Ciaro · 中文AI电影制作教程32

    How to evaluate the effectiveness of AI image generators in creating movie concept visuals

    When evaluating AI image generators for creating movie concept visuals, each tool should be given the same prompt, reference image, and number of attempts, and conditions such as tool version and settings should be recorded. The evaluation should first determine whether the still frame clearly presents the character's position, entry space, and door opening, then test lighting variations, and distinguish the roles of the reference image and the prompt. This method aims to reveal the generator's practicality for this type of shot, rather than just looking at whether the image is beautiful.

  4. note · miyuu_2026 · 图像视频创作分享44

    11/17 Making an AI Short Film with Wan2.2 + Krea2: Studio Shooting and My Daughter's School Entrance Ceremony, Part 2

    The author used Krea2 to generate images, Wan2.2 14B to generate video, and Capcut to edit a 15-second AI short film. The video was generated on Windows 11 with an RTX 5060 Ti 8GB and ComfyUI, at a resolution of 1024x576px and 73 frames, and edited as 3 seconds × 5 shots. The author says the short film did not appear in the YouTube Shorts feed.

  5. note · synth_brain · 图像视频创作分享52

    Two Ways to Connect ComfyUI and AI Agents: How to Choose Between MCP and REST API+SKILL

    The article compares MCP and REST API+SKILL, two ways for AI agents to operate ComfyUI. MCP uses fixed tool calls, offering high stability and controllable permissions, but it has higher installation and maintenance costs and is difficult to extend to the latest custom nodes. REST API+SKILL does not require MCP configuration and is suitable for quickly updated nodes such as Google Colab temporary URLs and MiniMax H3, as well as audio post-processing, but it requires SKILL.md to constrain the steps and command execution permissions.

  6. Reddit comfyui · 创作与工作流分享55

    Single RTX 3090 Real-World Test: Daily Usage Data for Local Agents and Image/Video Generation

    The author used a single RTX 3090 24GB, swapping VRAM in and out, to run a local agent and ComfyUI image/video generation simultaneously. Measured: LLM about 72 tok/s, SDXL image about 32 seconds, Wan 2.2 video 4-6 minutes. Data comes from 205 real requests and 35 rendering tasks, not synthetic benchmarks. The author concludes that 24GB can only accommodate either the LLM or one render at a time, and recommends using parallel 1 instead of 4.

  7. Post Magazine67

    Adobe Max Conference: AI Takes Center Stage.

    At the Adobe Max conference, Adobe announced new agents and tools for Firefly, Creative Cloud, and GenStudio, including Creative Agent, MCP integration, and Premiere generative media tools. The company also deepened partnerships with Google and Microsoft and showcased super-resolution tools from the Topaz Labs acquisition.

    Why it matters: Adobe launches multiple AI tools in Creative Cloud and Premiere and opens MCP integration, allowing creators to evaluate the impact of new features on existing workflows.

  8. Reddit StableDiffusion · 创作者与工作流分享13

    MOSS Image How to truly use reference images as visual input

    MOSS Image Users attempt to use reference images as visual input via the CLIP text encoder, with the model using Qwen 3.5 9B as a prompt enhancer; this method occasionally reproduces elements such as clothing from the input image, but is mostly completely ineffective. Users are seeking a method for consistent implementation.

Nov 16Mon
  1. note · akb428 · 图像视频创作分享58

    MiniMax-H3-Image Hands-On Test: Can Generate Character Art and Japanese Magazine Covers, with Pixubus Workflow

    The author tested MiniMax-H3-Image on a PC with 16GB VRAM, verifying that it can generate magazine covers with Japanese titles and subtitles, as well as character illustrations. Compared with P-H3-Image, the files are larger, but the Japanese text is more accurate; however, the hiragana 'ぱ' becomes 'ば', and the background sometimes turns black. A 4096×4096 image takes about 38 seconds to generate, but Pixubus does not currently support large-size output.

  2. Reddit StableDiffusion · 创作者与工作流分享16

    Minimax H3 Image Subject LoRA: Has anyone achieved good similarity?

    Training subject LoRA with Minimax H3 Image yields poor similarity, while the same dataset performs excellently on Qwen 2.1, Krea 2, and Moss. Increasing training epochs or steps showed no improvement; the Minimax H3 video model actually learns subjects better on static images. Motion learning is also difficult, and the poster is seeking advice on training settings for high-fidelity similarity.

  3. note · miyuu_2026 · 图像视频创作分享28

    11/16 [Wan2.2] Pink knit sweater girl street shooting episode

    The author used Krea2 to generate images, Wan2.2 14B to generate video, and Capcut for editing, producing an 18-second AI short drama of a girl in a pink knit sweater on a street shoot. 6 out of 8 shots were used. The production environment was Windows 11, RTX 5060 Ti 8GB, 32GB RAM, using ComfyUI workflow, resolution 1024x576px, 73 frames. The work consists of 3 seconds × 6 shots.

  4. Reddit StableDiffusion · 创作者与工作流分享57

    DiffusionBear v0.5.0 released: pure MLX runtime, virtual box, and Agentic REST API

    DiffusionBear v0.5.0 released, runtime changed to pure MLX and PyTorch removed, app size reduced from 1.3 GB to 615 MB, added virtual box, Kromoss 5B support, RGBA transparency, and Agentic REST API. Memory lifecycle changed to 3/4-phase sequential execution, total time for 220 benchmark tests decreased by 58.5%.

  5. Reddit StableDiffusion · 创作者与工作流分享56

    Minimax Image (Edit) Early Experience: Strong Concept-to-Photorealistic, Weak Multi-Reference Prompting

    Digital artist GrungeWerX tested Minimax Image (Edit) and found its concept-to-photorealistic capability outstanding, but multi-reference prompt following is weaker. He recommends starting at least at 16MP, with 30-50MP giving better results, and about 1 minute to generate 16MP on an RTX 3090 Ti.

  6. かみもと · 本地AI视频工作流72

    Hands-on test of the pDMD 2-step ComfyUI workflow for MiniMax-H3-Image: generation time on an RTX 4070 reduced by about 29%

    The pDMD 2-step LoRA for MiniMax-H3-Image, using ComfyUI 0.45.0 and an RTX 4070 12GB, reduced the average generation time for 9 image edits from 38.83 seconds to 27.62 seconds, a reduction of about 28.9%.

    Why it matters: The pDMD 2-step workflow reduced average generation time on an RTX 4070 from 38.83 seconds to 27.62 seconds. Composition and instruction following were largely maintained, but the lines became thicker.

  7. Aitrepreneur · YouTube 创作视频59

    MiniMax H3 Image Local ComfyUI Workflow Tutorial: 4K Generation and Multi-Image Editing

    MiniMax H3 Image can run locally in ComfyUI, supporting native 4K image generation, editing existing images, and combining up to 16 reference images. After a week of hands-on testing, the author provides one-click installation options for local and RunPod setups, covering text-to-image generation, multi-reference editing, inpainting, outpainting, and a built-in prompt enhancer.

  8. Reddit StableDiffusion · 创作者与工作流分享29

    Minimax H3 image model is criticized for being rushed to launch and having frequent issues.

    On the first day of its launch on ComfyUI, users reported that the Minimax H3 image model had issues: missing details at high resolution, abnormal anatomical structures, which were not fixed even with the p-h3 VAE; edit mode was even worse, lacking character consistency and being overly sensitive to prompts. Users suspect that Hailuo or the ComfyUI fine-tuned version has problems, and the released templates have also been weakened.

Nov 15Sun
  1. MindStudio Blog · 创作分享61

    MiniMax H3 Image Review: Text, Editing, and Realism Benchmark Tests

    MiniMax H3 Image performs well in text, editing, and realism evaluations, with large text, counting, and editing identity preservation performing well; small text and complex physics are weaknesses. The model is a 33B-parameter diffusion Transformer, 6-step sampling, native approximately 16 megapixel output, supports up to 16 reference images, local ComfyUI operation requires about 48GB VRAM. In editing tests, car swaps, outfit changes, and illustration-to-photo conversions all preserved composition and identity, but storyboard physics and motion details remain insufficient.

  2. Reddit comfyui · 创作与工作流分享12

    Qwen image edit 2.1 malfunction

    When creating character sheets in ComfyUI, Qwen image edit 2.1 cannot maintain consistent character facial features; characteristics such as head length, head width, and freckles change. The user tried about 100 prompts, using the base extension and without enabling the prompt enhancer, but the problem remains unsolved. Their hardware is a 5070Ti 16GB plus another card (ComfyUI does not merge) and 64GB RAM.

  3. かみもと · 本地AI视频工作流66

    MiniMax H3 Image Hands-on Test: 10 Style Conversions Using Only Prompts

    The author tested MiniMax H3 Image, using an illustration as the source image, and completed 10 style conversions including watercolor, oil paint, pen, and woodblock print using only prompts; after enabling the prompt expansion built into the ComfyUI template, the style changes became more pronounced, but the average generation time increased from 60.50 seconds to 187.12 seconds.

    Why it matters: The author tested MiniMax H3 Image, completing 10 style conversions using only prompts, and compared the effects and time costs of the prompt expansion toggle, providing a reference for local image stylization.

  4. Higgsfield 官方博客52

    Higgsfield becomes the most adopted image and video AI platform on Ramp

    Higgsfield became the most adopted platform in the image and video AI category in the Ramp Rate November 2026 ranking, with 42% of businesses that purchased such software through Ramp in the past 12 months choosing Higgsfield. Its adoption rate increased by 33 percentage points in a year, ranking first in six of seven rankings; Midjourney ranked second with 35%.

  5. SuSu_酥酥38

    Qwen 4.0 Max Preview's pelican test results are stunning.

    Qwen 4.0 Max Preview delivers stunning results in the latest pelican test, with significant improvements in anatomical details, riding posture, scene consistency, text rendering, and lighting and materials. Market expectations were previously low, but the latest test shows the model's detail and logical abilities appear to be optimized. The long beak, throat pouch, scaly feathers, and tail feather details are precise, the "PELICANO" text on the frame is clear and undistorted, and the overall cartoon three-dimensional feel is strong.

  6. SuSu_酥酥73

    MiniMax H3 Image quantized version reduces VRAM from 68GB to 6.4GB; 8GB graphics cards can run 4K image generation.

    The open-source community has quantized MiniMax H3 Image to Q2_K, reducing VRAM requirements from 68GB to 6.4GB, allowing 8GB graphics cards to run 4K image generation locally. The quantized version natively outputs 4096×4096 and 21:9 widescreen, official distillation produces images in 6 steps, integrates generation and editing, and supports up to 16 reference images.

    Why it matters: The quantized version reduces VRAM requirements from 68GB to 6.4GB, allowing 8GB graphics card users to run 4K image generation locally, which can be used to evaluate whether to replace existing image generation workflows.

  7. Reddit StableDiffusion · 创作者与工作流分享53

    H3 images rough? Three settings to fix it.

    The author shares three settings to fix H3 image roughness: increase size from default 4MP to 16MP (3552x4736), switch sampler to SamplerLCM with noise 0.9, and write a long prompt describing camera, lighting, and skin texture. The workflow file is in the first comment.

  8. Monster社区 · 中文ComfyUI实测与工作流54

    MiniMax H3 and MOSS Image Generation: Model Comparison and Accelerated LoRA Trade-offs

    Author T8star-Aix compares MiniMax H3 and MOSS image generation models, showcasing ComfyUI workflows and accelerated LoRA trade-offs. The article provides a checking sequence: establish a baseline with the same prompt, distinguish style from structure, then confirm whether the acceleration gains are worth the detail loss. It also notes that 6 steps, 15 steps, and 8 steps cannot be blindly copied without considering the model.

  9. note · akb428 · 图像视频创作分享29

    Create beautiful girl character parties with Anima and Pixubus 4: usable for novels, event CG, and animation

    Pixubus 4 and Anima can one-click generate a party of 1-5 beautiful girl characters, supports Japanese prompts, and no English tags are required. The party can be imported into Inkubus v4.5 for novel writing and output as EPUB, or via Vidubus 1.5 generate a 15-second animation with sound. Anima was developed by CircleStone Labs and Comfy Org, trained on millions of anime images and approximately 800,000 art images.

Nov 14Sat
  1. Reddit comfyui · 创作与工作流分享20

    MOSS Image excels in semi-realistic anime style.

    Reddit user Phenilethylamine showcased semi-realistic anime works generated using the MOSS-Image-SFT_clear model with LoRA. The workflow uses layer skip 22, interval 0-0.3, paired with two LoRAs: Mixdark Style and Realistic Background Enchanted, and upscaled with SEEDVR2.