comfyui-workflow-templates v0.11.99 released
comfyui-workflow-templates v0.11.99 has been released to PyPI and can be installed with pip install comfyui-workflow-templates==0.11.99.
comfyui-workflow-templates v0.11.99 has been released to PyPI and can be installed with pip install comfyui-workflow-templates==0.11.99.
The Green Flag project (44 shots) uses the on-the-one workflow to take a song in and produce a beat-locked music video, entirely on a single RTX 5080. The workflow uses Whisper word timings and a beat grid to generate a shot list. Each shot uses FLUX Klein to generate three keyframe options, which are then animated with LTX audio-to-video, and sound effects are generated. Finally, 1080p upscaling and frame-accurate audio alignment are performed.
The author open-sourced Director's Cut, a film and video production application based on ComfyUI that supports preset character creation and a video studio. The application uses Qwen-Image 2.1 to generate and edit images, MiniMax H3 to generate video with optional OmniRef, and also integrates an MCP server for AI assistants to call.
open-ai-canvas v1.7.3.1 released, optimizing short drama aspect ratio and video model selection: convert the work's aspect ratio based on model capabilities, fix the issue where 9:16 outputs as landscape; video models no longer force synchronized audio, gray out unavailable models based on aspect ratio; tighten fallback logic for storyboard duration, character/prop attachment, etc.
Reallusion releases a free update for Character Creator 5.2, integrating AI 3D model generation, character wrapping, auto-rigging, wearable clothing, and real-time animation into a unified workflow. The new version integrates Tripo and Hunyuan generation models, supports converting almost any humanoid mesh into CC native topology, and provides an AI credit system to manage 3D generation, motion capture, and AI image/video/audio services.
Why it matters: CC 5.2 integrates AI generation, wrapping, rigging, clothing, and animation into a single pipeline, allowing creators to evaluate whether to replace multi-tool workflows.
ifanr hands-on tested Jimeng Short Drama Studio and found that it integrates scripts, assets, storyboards, previsualization, and modifications into a complete short drama creation workflow. Among them, DreamScript handles scripts, Anycook generates characters and scenes, DreamShot breaks down storyboards, the 3D director's stage and intelligent previsualization validate the direction before generation, and segment reshoots and frame capture annotations support local modifications. Seedance 2.5's 720P with reference video generation is equivalent to as low as approximately 0.19 yuan per second.
Why it matters: Hands-on test of Jimeng Short Drama Studio's complete creation workflow, see how it uses asset reuse and intelligent previsualization to shift the cost-reduction logic of short dramas from cheap generation to fewer redo.
LibTV launches its annual major promotion for Double 11: Wan 3, Kling 4.0, and Minimax H3, three flagship video models, are available for unlimited use with 0 credits for up to 30 days. The Supreme 92K annual package is discounted by 24,000 yuan, with an event price of 16,999 yuan. The platform also offers 50+ exclusive features and 10+ creator monetization paths, with the viral hit support plan offering up to 100,000 yuan per single piece.
OpenKatana is an open-source video editor that can be driven by AI agents from ComfyUI, released by author rk0neki in Comfy's open-source Higgsfield challenge. It adds 85 nodes and six apps, based on the Comfy SDK, runs locally, Apache-2.0 licensed, no account required. In the demo, Claude completes editing through Studio, rendering engine, scripts, and skills; the generation model needs to be installed separately.
Why it matters: An open-source, locally running AI-driven video editing solution that creators can evaluate as a replacement for cloud tools.
ComfyUI-MOSS-QwenVL-Layout node v1.1.0 adds a prompt-only mode that can convert base prompts into complete bbox-guided prompts. The node uses Qwen3-VL / Qwen3.5-VL to detect objects and text in images, constructs MOSS-Image layout JSON prompts, and includes built-in image scaling. It also provides a change request box, supporting edits such as recoloring, clothing replacement, and adding objects.
Suno launches MCP Beta, AI agents can directly call to generate music without manually switching tools. Users can apply for access via the early access link.
The Freefield open-source project rebuilds the functionality of an AI creation website into ComfyUI workflows and automatically generates a Web UI. The repository currently has 120 application folders, 119 marked as ready, covering 95 of the 129 tools listed by Higgsfield. End-to-end measurements show that the morph application takes 160.8 seconds from button to video, mixed-media 149 seconds, and dub 72 seconds (H100).
Why it matters: Freefield automatically generates a Web UI from a workflow manifest, rebuilding 120 AI creation applications as ComfyUI workflows, providing creators with a reusable architectural reference.
Gemini 4 Argon fixed MOSS's BBox workflow, enabling it to support arbitrary aspect ratios and resolutions. Previously, the default workflow would experience coordinate offsets at resolutions above 1MP, and changing the aspect ratio would not update the drawing canvas. The coordinate fix applies only to MOSS, while the canvas fix applies to any workflow using that node.
Darius is a 100% locally hosted creative canvas that reimagines the ComfyUI workspace, supporting multiple image and video models on a single canvas. A local chat (ornith 1.5 9b) is available to discuss concepts and refine prompts. It supports full audio mode, saving character/story/shot ideas, and is accessible on mobile and LAN devices. On first run, it checks for and downloads missing dependencies and models. The repository is on GitHub.
Pika announces that Kling 4.0 is now available on its platform, supporting full reference, multiple keyframes, and generation up to 30 seconds. Pika says this is Kling's most powerful update in terms of control, allowing users to create scenes they envision in their minds.
Why it matters: Kling 4.0 launches on Pika, bringing full reference, multiple keyframes, and generation up to 30 seconds, allowing creators to evaluate new scene-building workflows.
patch-gg open-sourced anymove, a workflow that converts AI videos into 3D animations, supporting creatures such as dragons, spiders, wolves, and demons, with a cost of about $0.26 per clip. The method uses dual-camera same-frame capture plus joint reader fitting to generate GLB files at 24fps, requiring no manual editing and completing within one minute. The repository also includes Agent Skills and MiniMax H3 LoRA.
Why it matters: This open-source workflow converts AI videos into 3D animations, using dual-camera same-frame capture plus joint reader fitting, with a cost of about $0.26 per clip, and can be transferred to other creature characters.
The author trained a retro anime style LoRA, recreating the cel texture, film grain, and hand-painted background style of 80s/90s OVAs. Recommended weight is 0.7-1.0, trigger word is retro anime style illustration, and it can be downloaded at https://civitai.red/models/1970260/retro-anime-style.
HeyGen Voice launches a speech-to-speech (Speech-to-Speech) feature. Users can record a line with their phone or computer and get studio-quality voiceover that preserves emotion, rhythm, stress, and pronunciation, changing only the voice itself. The feature is now available on HeyGen and the API, supporting clean narration from casual recordings and offering a richer emotional range than text prompts, including whispers, laughter, and accents.
Why it matters: HeyGen's speech-to-speech turns phone recordings directly into studio-quality voiceover, reducing repeated TTS generation, and creators can use it to evaluate the new voiceover workflow.
Foundry releases Nuke 18.0 open beta, adding a Griptape node that lets compositors run generative AI image and video workflows directly in the Nuke node graph. Nuke Studio adds automatic cut detection and annotation tool updates, 3D Gaussian Splats workflows are also improved, and support for RED cameras, Nikon N-RAW, and H.265/HEVC media.
Developer ItsBalanse has developed try2create, a one-click image, video, and audio generation application with ComfyUI as the backend. Users only need to press a button to generate, without dealing with the underlying workflow.
FireCut is an AI editing assistant for Adobe Premiere Pro and DaVinci Resolve that automatically scans timelines, removes silent segments, deletes duplicate shots, and generates multilingual subtitles. Its Magic Cut feature lets you rough-cut raw footage into a first draft with a single click, no script required. FireCut is now available for a free trial.
DaVinci Resolve 21.1 integrates AI assistants Claude, Claude Code, and ChatGPT Codex at IBC2026, allowing users to analyze projects, organize media, adjust settings, and batch render using natural language.
Why it matters: AI assistants lower the barrier to Resolve script automation to natural language, allowing editors and colorists to complete repetitive tasks such as metadata organization and batch rendering more quickly.
IronLeo78 released ComfyUI custom node IronLeo78 MTV v0.0.2 for music video generation. The node is based on MiniMax H3, automatically segments by song length, and after audio separation, makes the visuals follow a single instrument, and supports continuity guidance carried over from the previous frame. The author is a music video editor with 20 years of experience. The tool is available in the ComfyUI custom node manager, and the code is on GitHub.
FrameForge is an open-source desktop application submitted by the author to the ComfyUI Higgsfield Build Competition, focused on building movies with Minimax H3.
ComfyUI creator crinklypaper releases H3 Relay Hires node for MiniMax H3 anime videos: first render a preview at 0.5 MP, upscale via AnimeSR to 1920, then use one-step H3 turbo to redraw details; a 5-second 1920x1088 clip takes about 118 seconds end-to-end, about 4x faster than native ~500 seconds, and no low-step LoRA is needed.
Jimeng launches a short drama studio feature, linking script, assets, storyboards, generation, and post-production editing into a single project. In testing, the Agent organizes character appearances across different periods based on the plot, maps characters, scenes, and prompts to specific storyboards, and generates transition videos to fix segment connection issues.
Why it matters: Jimeng Short Drama Studio launches, linking script, assets, storyboards, generation, and post-production into a single project, providing a reference organizational approach for AI short drama creation.
Apple Creator Studio update: Pixelmator Pro adds Generate Fill, which can rebuild selected areas of an image based on text descriptions. Final Cut Pro includes MotionVFX effects, Logic Pro adds an AI assistant based on Apple's own models and OpenAI, and the update will be pushed for free in November.
comfyui-workflow-templates v0.11.98 has been released to PyPI, adding Patina and VEED Dubbing templates. It can be installed via pip install comfyui-workflow-templates==0.11.98.
ComfyUI v0.43.0 released, adding MOSS-Image RoPE Scaling support, color depth control, and MOSS-Image-Edit LoRA support; also adding Fibo 2.0 support, changing its RoPE Scaling to NTK-aware and fixing text-to-image negative prompt handling, along with other fixes and improvements.
If you can describe it, you can run it as a video ad. Runway Ads turns your creative direction into brand ads tailored for every platform, then learns what works and produces ads that go beyond it. Teams like @Lovable are already using Runway Ads to scale ad creative and shorten production time.
Fylmer is seeking early testers. Its built-in model Tresca is based on MiniMax H3, and it offers free credits for testing. Testers can bring their own API keys and use other models via MCP connections. Fylmer is a filmmaking workspace that integrates scripts, characters, references, and shots.
Creator Devajyoti1231 released the open-source project Openjutsu, implementing local one-click video character replacement based on MiniMax H3. Upload a video and a character image to run it; it replaces segment by segment while preserving the original audio, and also supports modes such as object replacement, background modification, generation from images, face restoration, and color restoration.
Jimeng AI launches 'Short Drama Studio', integrating script organization, character and scene management, storyboard generation, video production, and post-production adjustments into a single platform. Some features integrate the same capabilities as Xiaoyunque Short Drama Agent, and it is equipped with the first professional script model DreamScript, supporting upload of 100,000-character long texts and continuous generation of 100 episodes, while also launching team edition management capabilities.
Why it matters: Jimeng integrates the full-process tools for short dramas into a single platform and reuses Xiaoyunque Short Drama Agent capabilities, providing creators with a one-stop alternative and reducing the cost of switching between multiple platforms.
Kling 4.0 was opened to all users on October 15, with official showcases of its application cases in film, advertising, and animation production. The Fountain 0 team, based on early testing, expects discarded footage to be reduced by at least 50%, and animation director Cao Han mentioned more accurate execution of detailed creative instructions.
A batch of open-source AI short-drama workbench projects have appeared on GitHub, covering the entire workflow from script, storyboard, generation to editing. The article summarizes seven projects: Huobao Drama uses four AI Agents for division of labor, Toonflow supports infinite canvas and 3D director's console, and Waoowaoo received nearly 3,000 stars on its first day of open-sourcing.
Why it matters: A roundup of seven open-source short-drama workbenches, covering the full workflow from script, storyboard, generation to editing, each with its own focus, so creators can choose according to their needs.
Runway launches App Studio, allowing users to describe their needs in natural language and build custom creative apps with standalone interfaces without code, generating images, videos, and audio. App Studio includes built-in apps such as Character Swap, AutoComp, and Product Portfolio, built on Brand Kits, Workflows, and Assets. It is included in all paid plans and consumes existing Runway credits.
Why it matters: Runway's App Studio brings no-code app building into creative work, allowing teams to turn years of accumulated workflows and brand standards into tools anyone can use, reducing reliance on individual experts.
infinite-creation v1.0.2 released, fixing some bugs and concurrency issues with RunningHub API tasks. The Agent now has internet search capability. It supports searching and downloading reference images and generating assets via image-to-image, automatically searching for materials to fill plot gaps, and can generate storyboard videos using 'image-to-video with first/last frames / text-to-video'.
infinite-creation v1.0.3 released, fixing some bugs and optimizing the categorization of the configuration page. It now supports generating multiple candidate videos from a single storyboard, allowing users to choose the best one for export (equivalent to a gacha/draw feature).
Topview Film Studio 2.0's Cinematic Suite currently has 25+ tools, and the author highlights five of them.
mlx-audio v0.6.2 released, adding Ming 3.0 TTS/ASR/Audio Tiny models, Granite-Speech-5.0-2B ASR/AST models, and Phonon 2.5 STT models.
ComfyUI creation project Liquid Inference releases an update, adding local account login, price display, and image cost features. It also fixes the issue where example templates could not run, and adjusts model naming in Registry text.