Skip to content
Date and source

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

Oct 6Tue
  1. あいきみ · 图像编辑与LoRA实测58

    RTX 3090 Hands-On with MiniMax H3's 360° Orbit LoRA: Generating Orbit Videos from a Single Image

    The author tested pablodawson's MiniMax H3 360° Orbit LoRA on an RTX 3090, using the same image as both the first and last frame of the video, generating an approximately 3-second, 73-frame orbit video. The back of the subject, not captured in the original image, appears midway through.

  2. note · life_good_life · 图像视频创作分享22

    2026 AI Image Generation Tool Use-Case Quick Reference: Recommendations by Use Case and 20 Copy-Paste-Ready Prompts

    This 2026 quick reference provides AI image generation tool recommendations by use case: choose Midjourney for visual aesthetics, ChatGPT or Gemini for Japanese text in images or conversational editing, and Adobe Firefly for commercial images where copyright compliance matters.

  3. note · lofty_eagle5366 · 图像视频创作分享12

    Generating Images with 150,000 PixAI Points Saved Up

    The author spent all 150,000 points saved in their PixAI account on image generation in one go, with none of the works using reference images. The generation process is shown in order, with one failed image costing 7,000 points while the daily login reward is 10,000 points. Since relying solely on prompts had limited results, the author spent another 100,000 points training a LoRA model, feeding it 10 semi-realistic images they had on hand, and felt the results improved noticeably.

  4. ArchiGen AI · 图像生成与ComfyUI实战22

    Giving Renderers a Normal Bad Input: A Weak-Input Benchmarking Method for AI Renderers

    ArchiGen proposes testing AI renderers using the weakest source images a studio can normally accept, fixing the camera and key design facts, and performing only one initial generation and one controlled revision. The method requires producing five frames: an acceptable source image, default first pass, controlled fix, delivery export, and a difference overlay, scored by the number of changed invariants, first usable output, and fix time. The article cites Rendero's comparison of 11 AI rendering products, noting that Rendero did not actually execute this test.

  5. note · lovely_hebe1274 · 图像视频创作分享22

    How to Request AI Portrait Corrections? Turning 'Just Optimize It a Bit More' into Specific Instructions

    For the problem of wanting to fine-tune AI-generated images but not knowing how to describe it, a correction instruction method suggests first specifying the base image and locking specific locations, then pairing 'what's unsatisfactory' with 'what you want it changed to,' for example requesting 'the character on the right side of the image closes their mouth with a gentle expression, corners of the mouth slightly raised, no big laugh.' The article provides an input template with fields for use case, top priority change, desired effect, parts to preserve, and changes to avoid, and recommends changing only the most important part at a time and comparing the corrected version side by side with the original.

  6. GitHub:AI视频开源项目发现65

    mushigaite open-sources short-video-maker: automatically cuts YouTube long videos into vertical short videos

    mushigaite has open-sourced short-video-maker on GitHub, which can automatically cut YouTube videos or local files into 9:16 vertical short videos, integrating highlight detection, subtitles, translation and voiceover, with no intro credits or watermarks.

    Why it matters: The open-source project strings long-video slicing, subtitles, translation and voiceover into a single local pipeline, and can serve as a free alternative reference for short-drama overseas editing tools.

  7. 阡陌的小站 · 本地创作实战70

    Tutorial: deploying Qwen-Image-2.1 and MiniMax H3 locally in ComfyUI with 8GB VRAM

    Qianmo published a tutorial on generating images and videos locally via ComfyUI on an RTX 4070 Laptop with about 8GB VRAM, using the Qwen-Image-2.1 jailbreak GGUF workflow for images and MiniMax H3's 04_i2v_fl2v_4step_sla_lowvram workflow for videos.

    Why it matters: A complete path to running Qwen-Image-2.1 and MiniMax H3 under 8GB VRAM, including model files, directory structure and parameter configuration, reproducible by comparison.

Oct 5Mon
  1. CoAnimator · 可编辑视频与动画工作流22

    The 7 levels of Agentic AI video creation: from manual editing to agents maintaining projects

    CoAnimator proposes a 7-level framework for Agentic AI video creation, from manual editing, AI-assisted tasks, prompt generation, and predefined workflows, to level 5 where the agent leads editing, level 6 where the agent checks and revises, and level 7 where the agent maintains saved projects. The framework distinguishes workflows from agents across three dimensions: decision authority, result checking, and maintainability, and uses a 40-second product demo update as an example to illustrate the differences in responsibility at each level.

  2. note · hagiri1312 · 图像视频创作分享22

    Fixing the same character with AI portraits: a method for choosing which features to keep in a trial of note's mascot character expressions

    Japanese creator hagiri used the September 29 character design sheet as a baseline to trial 5 expressions and a thumbnail for note's mascot character with AI, by separately specifying "features to keep" and "features that can vary." The prioritized fixed features are a soft face shape, relatively large brown eyes, straight bangs and shoulder-length fluffy brown hair, a white shirt with a light beige-to-pink cardigan, and a soft anime style, while expression, pose, props, and background change with the article scene. The author emphasizes that this was only a small trial, does not guarantee generating the same face in one go, and did not measure time saved.

Oct 4Sun
  1. note · hideout927 · 图像视频创作分享40

    Generating character design sheets with MiniMax H3 in ComfyUI: no dedicated LoRA needed

    MiniMax H3 can leverage its own reference image and prompt comprehension capabilities to generate character design sheets in ComfyUI without a dedicated LoRA or additional training; generation uses Core nodes, and custom nodes only handle generating JSON and natural-language prompts. The node combines view instructions such as front, side, back, and facial close-up with the original image to keep the same character and clothing consistent, with H3 filling in parts not photographed. Japanese prompts perform worse, while English prompts pass more easily; if separate hand and foot images are not needed, this feature can be turned off.

  2. note · aoao_ai · 图像视频创作分享50

    AI comic backgrounds always inconsistent? A practical method for fixing room layout with a "background setting sheet"

    When drawing manga with AI, the positions of doors and windows, furniture layouts, and room sizes change every time the same "home" or "nursery" is generated. To address this, the author created a "background settings sheet" independent of the character settings sheet, recording basic layouts such as location type, door/window and large furniture positions, and room relationships, along with composition and description rules—for example, never using a frontal composition for the entryway, fixing the refrigerator position, and fixing equipment like the nursery slide. Small items and items that appear only in a given episode are not included in this sheet and are managed on the script side; the settings sheet is fed directly to the image generation AI rather than being rewritten into prompts each time.

  3. note · lucky_allium8251 · 图像视频创作分享31

    DALL-E 3 Doesn't Read Japanese Directly: It Internally Translates to English First, Then Generates Images

    Independent verification by UX designer Yennie Jun shows that DALL-E 3 does not directly process Japanese prompts; instead, it first internally translates and expands them into English before passing them to the image generation layer. The verification covered 6 languages, with 20 images per prompt for a total of 120 images, and found that Zulu-language prompts generated African-descended characters at a rate exceeding 50%, while English prompts generated them at only about 20%.

  4. DesignerBox · 商品图与视频工作流58

    How to Use AI Music for Advertising Jingles: A Five-Step Process and a Comparison of Commercial Terms Across Tools

    Creating an advertising jingle with AI music can be broken down into five steps: write lyrics, generate and lock the song, measure the word-by-word timeline, create visuals by section, and add subtitles. The author compares licensing terms across platforms, noting that Suno only transfers rights of generated results to users on its paid tier, Udio's current terms are limited to personal non-commercial use with downloads disabled, and ElevenLabs' self-service plans exclude film, television, and broadcasting. Meta's and TikTok's general music libraries are both unavailable for commercial advertising. The U.S. Copyright Office considers purely AI-generated content ineligible for copyright protection, with human-written lyrics being the most defensible component for asserting rights.

  5. DesignerBox · 商品图与视频工作流22

    Ad Script Examples: 8 Video Scripts with Word Counts

    At a reading speed of 183 words per minute, a 15-second ad script is approximately 35 words, a 30-second script about 70 words, and a 60-second script about 145 words, leaving roughly 20% of the time for visuals without narration. The article provides 8 script examples for 15/30/60 seconds, with each line divided into "Narration" and "Visuals" columns. It also explains that when writing for AI voiceover or AI hosts, each sentence should be one breath, kept to approximately 15 words or fewer, and generated hosts must not claim to have used the product.

  6. CoAnimator · 可编辑视频与动画工作流34

    What Is the Agentic AI Video Tech Stack? Models, Agents, Tools, and Workflows

    Agentic AI video consists of replaceable layers including models, agents, skills, MCP/tools, projects, renderers, feedback loops, and workflows, with each layer's failure corresponding to different repair methods. Models handle reasoning but cannot operate on project files; agents (such as Claude Code and Codex) can read and write files, run commands, and view their own rendered results.

  7. ArchiGen AI · 图像生成与ComfyUI实战38

    ComfyUI Architectural Rendering Tutorial: Lock Lighting First, Then Fix Glass Reflections

    The correct order for fixing reflections in ComfyUI architectural rendering is to first lock the final lighting, then use a glass mask combined with depth or Canny guidance to repaint only the glass areas at low denoise. The official ControlNet examples note that the Apply node does not automatically generate depth or Canny maps—users must provide matching inputs themselves. The official inpaint workflow also only repaints the selected area. The author uses five checkpoints for quality control and finally confirms with a difference mask that changes have not exceeded the mask boundaries.

  8. Aitrepreneur · YouTube 创作视频62

    Aitrepreneur Releases a ComfyUI Multi-Image Editing Workflow for Qwen Image 2.1

    Aitrepreneur showcased a custom Qwen Image 2.1 ULTRA ComfyUI workflow supporting multi-image editing with up to 9 images, allowing users to replace characters, remove or move objects via text prompts. The workflow also includes operations such as inpainting, outpainting, and background removal, and is used in conjunction with additional tools and LoRAs.

  9. note · yoseiya · 图像视频创作分享22

    Making AI Videos with a Phone: From "I Want to Make a Video Like This" to a Finished Work

    Creator Youkiboya started with the concept of "taking a walk through a colorful Japanese garden," using ChatGPT to complete the ideation, storyboarding, and still frames, then converting them into video and editing the final cut, with music generated by SUNO. The entire workflow is phone-centric, designed for beginners unfamiliar with complex prompts. This note is currently priced at 980 yen, and future updates and revisions may adjust the price.

  10. 小智博客 · 音视频实践38

    MoneyPrinterTurbo hands-on test: can AI video generation get local production, remote access, and license verification working?

    The hands-on test examines whether MoneyPrinterTurbo can truly run the entire pipeline of "topic → script → assets → voiceover → subtitles → finished video," covering three stages: local production, remote access, and license verification. The original article lays out a fairly complete test path; the focus is not the broad question of "whether AI can make videos automatically," but the practical usability of the tool's end-to-end workflow.

  11. 小智博客 · 音视频实践38

    Don't want to learn React first? Use Codex to help you finish your first Remotion motion-graphics video

    With Codex's assistance, you can complete your first Remotion motion-graphics video even without being familiar with React. Remotion writes video frames as React components, with animation, layout, and data all controlled by code, skipping the traditional workflow of cutting assets on a timeline, setting keyframes, adjusting transitions, and then waiting for rendering.

Oct 3Sat
  1. note · ai_to_hibi_note · 图像视频创作分享36

    Writing 8 short stories with local AI on an ordinary laptop and turning them into narrated videos: a hands-on test of Qwen3.5, FLUX.2 klein, and VOICEVOX

    On an ordinary laptop without a GPU (Ryzen 7 7735HS, 32GB RAM), local AI was used to have Qwen3.5 (4B) write stories, FLUX.2 klein 4B draw illustrations, and VOICEVOX provide voiceover, producing a 40.9-second vertical narrated video with 6 scenes.

  2. note · sakisuta · 图像视频创作分享34

    How to choose between GPT Image 2.5's Flare and Sunburst models? A summary of 6 editing features

    OpenAI's GPT Image 2.5 includes two built-in models, Flare and Sunburst: Flare is the default model, with higher image quality than GPT Image 2.0 and 2–4 times the speed; Sunburst is the highest-precision model, slower to generate but with finer control, producing more realistic clothing materials and photographic texture under the same prompt.