Skip to content
Date and source

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

Sep 24Thu
  1. 公众号:短剧内行人20

    Xiaowu Brothers' "Global Spark Plan": How AI Short Dramas Going Overseas Can Bridge the Monetization Gap

    On September 17, Xiaowu Brothers launched the "Global Spark Plan" in Hangzhou, partnering with tool partners such as Wondershare Drama Factory, Vidu, SenseTime Seko, AniShort, and Nano AI to connect topic selection, production, translation, rights confirmation, and attribution-based distribution, attempting to solve the monetization gap in AI short dramas going overseas where "skills keep growing but not a cent is earned abroad." The plan relies on a distribution network covering 180 countries and regions and over 8,000 overseas channels, allowing works to first be tested by real overseas audiences, then using monetization data to guide production.

Sep 23Wed
  1. Adobe · 官方创作与影视博客62

    Adobe completes acquisition of Topaz Labs, integrating AI image quality enhancement into Firefly and Photoshop

    Adobe announced the completion of its acquisition of AI video and image enhancement company Topaz Labs, which will retain its independent brand, with its applications and models continuing to be offered as standalone products. Topaz Labs' technology has been integrated into Adobe Firefly and Photoshop for super-resolution and enhancement of images and videos, with its Neurostream technology supporting processing of complex AI models on local devices or Topaz's cloud.

    Why it matters: Adobe integrates Topaz Labs' image quality enhancement models into Firefly and Photoshop, offering insight into the integration trends of post-production toolchains.

  2. Loomance · 本地AI动画实测34

    How to turn artwork into animation locally: A hands-on comparison of Loomance and ComfyUI on WAN 2.2

    On an RTX 5070 12GB GPU, both ComfyUI and Loomance can run the WAN 2.2 image-to-video model entirely locally, without cloud rendering or per-use fees. ComfyUI offers node-level control but requires manual configuration, while Loomance wraps the same model into three preset workflows: Stable Scene, First → Last Frame, and Wild Mode, allowing direct rendering from children's hand-drawn sketches as source images.

  3. Wireflow · AI短剧与视频制作教程30

    Best Wan 2.5 API tools of 2026: A roundup

    Eight API platforms integrating Alibaba's Wan 2.5 video generation model are compared one by one in 2026, covering price, resolution, duration, and audio output. Wan 2.5 can output dialogue, ambient sound, and background music in a single generation, supporting 720p/1080p and 5 or 10 second clips. Kie.ai is the cheapest option at about $0.06/second for 720p and about $0.10/second for 1080p, fal.ai offers full parameters at $0.15/second for 1080p, and Alibaba Cloud Bailian is the first-party channel.

  4. Wireflow · AI短剧与视频制作教程66

    How to Create AI Song Ads on a Single Wireflow Canvas Without CapCut

    Wireflow publishes a tutorial on how to complete an 80-second AI music ad on a single canvas — from script, song, and word-level timeline to stills, motion, subtitles, and end credits — without ever exporting to CapCut or Premiere.

    Why it matters: A complete breakdown of the eight steps to produce an 80-second AI music ad, from writing lyrics to rendering, along with the actual cost of approximately 6,100 credits and a record of pitfalls encountered.

  5. 阿真Irene · 创作者专栏78

    Hands-on with Vidu S2: Case Breakdown of the Real-Time Interaction and Real-Time Editing Model Lines

    Vidu S2 was released on September 15, with the paper published on September 10. The web version, app, mini-program, and API are all open for use, featuring two models: real-time interaction and real-time editing. The author tested the real-time interaction model S2-Avatar, which supports voice conversation, real-time motion response, and reference image upload for interaction, and can save 720p interaction videos. Tests included virtual idol livestreaming, selling Cantonese morning tea, English conversation, and full-body scenes; after being interrupted mid-conversation, the character could pick up new topics.

    Why it matters: The author uses extensive hands-on cases to break down Vidu S2's two capability lines — real-time interaction and real-time editing — providing a basis for assessing the viability of digital human livestreaming and real-time outfit-change scenarios.

Sep 22Tue
  1. note · kentauber_ai · 图像视频创作分享36

    A Practical Guide to Creating Anthropomorphized Yokai Runway Videos with MiniMax H3

    A creator used MiniMax H3 to anthropomorphize Kumo-jiko (spider woman), Yuki-onna (snow woman), Kara-tengu (raven tengu), and Kyūbi (nine-tailed fox), producing Japanese-style runway videos. Their experience: inputting full-body images and face close-ups as Picture 1 and Picture 2 respectively can reduce facial distortion; setting a background makes the visuals more stable but viewers lack a sense of dynamism; fixed poses still leave room for more creative expression.

  2. 小智博客 · 音视频实践60

    Hands-on Environment Tuning for Running MiniMax-H3 on ComfyUI with a Single Hygon K100_AI Card

    The author documents the hands-on process of tuning and running MiniMax-H3 on ComfyUI with a single Hygon DCU K100_AI card. MiniMax-H3 is an open-source 33B omni-modal video generation model by MiniMax, supporting three modes: text-to-video, image-to-video, and reference-to-video. It natively outputs 2K/15-second videos with stereo audio. Hygon's official solution recommends 144GB of VRAM per card.

  3. VibeClips · AI视频创作教程22

    Why Viral AI Video Clips Go Viral: Five Formats, the First-Second Hook, and a Model Selection Guide

    Viral AI video clips succeed through a hook in the first second, formats built around a single surprise point, and choosing the right model for a specific effect — not hashtags or trending music. Most shared clips are 5 to 12 seconds long, and TikTok officially recommends placing the hook within the first three seconds.

  4. Wireflow · AI短剧与视频制作教程22

    2026 Best White-Label AI Video Software Tools Roundup: A Comparison of Eight Tools Including Wireflow, Banuba, and Creatomate

    Wireflow ranked first among 2026 white-label AI video software tools. It chains script, image, video generation, and compositing into a node-based pipeline triggered via its own API, so clients can receive unbranded finished products without ever logging into the canvas.

Sep 21Mon
  1. 腾讯 ARC · 创作模型60

    Tencent ARC Releases WorldCrafter-Fast Video Generation Weights

    Tencent ARC released self-contained WorldCrafter-Fast weights on Hugging Face, supporting image-to-video, text-to-video, and an interactive image-to-video demo, with inference possible without WorldCrafter-Base files.

    Why it matters: WorldCrafter-Fast has released standalone weights, with sampling step configurations for both image-to-video and text-to-video made public, making it easier for creators to evaluate integration costs.

  2. GitHub:AI视频开源项目发现62

    Tencent Hunyuan Open-Sources Prism: Dynamic Sparse Attention for Native 2K Joint Video-Audio Generation

    Tencent Hunyuan, in collaboration with Fudan University and Zhejiang University, proposed Prism, a dynamic sparse attention framework for natively training 2K joint video-audio generation models. It organizes token sequences into spatiotemporal macro-blocks and dynamically allocates block shapes based on local content, achieving a 2.5x training speed improvement over full attention with higher generation quality.

  3. 公众号:短剧深水区38

    How Does a 3-Year-Old AI Cute Baby Beat Cultivation Grandmasters? Breaking Down the Cute Baby Genre with "Listening Water Little Koi" and "Nuonuo Descends the Mountain"

    The AI realistic-human short drama "Listening Water Little Koi" accumulated 1.2 billion views on Douyin's paid tier in 4 days, with a single-day peak of 700 million, and 56.01 million heat on Hongguo's free tier. The AI comic drama "Nuonuo Descends the Mountain, All the Senior Brothers Are Panicking" surpassed 1 billion Douyin views after launching on August 25, with over 10 million pre-orders on Hongguo before Season 2 went live.

  4. Clip设计夹 · 创作教程与项目分享36

    Meitu Design Studio Uses AI to Batch-Test Styles from a Single Image, Handling the Full Set of Materials for Autumn Collection Launches

    Meitu Design Studio launched a clothing try-on feature that lets users upload a single flat-lay product image to batch-generate try-on renders with different models, poses, and scenes. It also supports AI clothing color change, outfit replacement, and model and scene switching. Users can also generate a full set of images with a single sentence in the homepage dialog box, or upload a competitor's hit image to replicate its lighting and composition 1:1, then continue generating sales videos with scripts automatically produced by an agent team. The tool has been among the first to integrate with mainstream large models such as WorkBuddy and Codex, and supports mobile app operation.

Sep 20Sun
  1. ComfyUI-Easy-Media · 开源创作项目发布22

    ComfyUI-Easy-Media v1.3.2 Released: New MiniMax H3 Progressive SelfLift Sampling

    ComfyUI-Easy-Media released v1.3.2, replacing character context exchange with pure Drift-Control AV and adding progressive SelfLift sampling for MiniMax H3. This version also improves the multi-track editor UI and universal tiling, and fixes issues with video reference frame and audio-locked video frame alignment, context latent selection, and duplicate semantic tags in prompts.

Sep 19Sat
  1. Pika60

    Pika releases new creation platform, Relight Media supports adjusting video lighting

    Pika has released a new AI creation platform, which the company says is aimed at creators and pursues output that meets strict standards. Using Relight Media as an example, the company explains that beyond changing the time in a video, it can also adjust the color, intensity and direction of lighting, and invites users to try it on the new Pika.

  2. Pika65

    Pika launches Camera Director, enabling multi-angle reshoots of any clip

    Pika releases a new version of its product and the Camera Director app: upload one video to generate a whole set of replacement shots while preserving the original clip's motion and performance. The company calls the new Pika an AI creation platform for creators, emphasizing simplicity, comprehensiveness and output aimed at high standards.

    Why it matters: Pika officially demonstrates the multi-angle reshoot feature, which creators can use to judge whether existing footage can directly generate replacement shots.

  3. Wireflow · AI短剧与视频制作教程22

    Best product video generators of 2026: a review of eight tools including Wireflow, Creatify and PixVerse

    The 2026 product video generator review ranks eight tools on four criteria: product fidelity, reusability, shot control and output aspect ratio. Wireflow ranks first, linking product images, video models and an editor on a node canvas so the same workflow can be reused in bulk across hundreds of SKUs; Creatify generates AI presenter ads directly from a product link, PixVerse excels at camera movement in image-to-video, and HeyGen is known for lip-sync and multiple languages.

  4. Pika60

    Pika launches Product Ad app to generate product ad films from brand assets

    Pika has launched the Product Ad app, aimed at situations that need high-end product videos but cannot afford high production costs. The app reads the user's strategy brief and brand assets, confirms the script and storyboard together with the user, and ultimately produces a video ready to run. Pika also says the new Pika is an AI creation platform for creators, of which Product Ad is only one part.

Sep 18Fri
  1. GitHub:AI视频开源项目发现70

    Alibaba Open-Sources SparkDiffusion: A Video Generation Acceleration Framework for Wan 2.1/2.2

    Alibaba open-sourced SparkDiffusion, an acceleration framework for DiT video generation models that combines sparse low-rank attention RoLa, few-step distillation CrossDistill, and self-developed high-performance operators, achieving over 200x end-to-end inference speedup compared with dense multi-step baselines.

    Why it matters: Alibaba open-sourced SparkDiffusion, using sparse attention and few-step distillation to speed up Wan 2.1/2.2 inference by more than 200x, and released multiple weight tiers.