comfyui-workflow-templates v0.11.106 released
comfyui-workflow-templates v0.11.106 released to PyPI, adding LTX VFX API templates (reframe, upscale, SDR to HDR) and MiniMax-H3-Image partner API templates.
comfyui-workflow-templates v0.11.106 released to PyPI, adding LTX VFX API templates (reframe, upscale, SDR to HDR) and MiniMax-H3-Image partner API templates.
Adobe releases After Effects 27.0, adding an agentic AI assistant that can control the software via text prompts and automatically generate JavaScript expressions. The new version also adds the Contour Paint effect, 3D material options, and improvements to import/export and auto-save; the AI assistant's creative operations consume generation credits, with single-app subscriptions providing 25 free credits per month.
Why it matters: The new version adds an agentic AI assistant that can automate tasks via text prompts and generate expressions, allowing creators to evaluate its credit consumption and workflow fit.
Soundly partners with Avid to launch a free Pro Tools extension that brings its 15-year-accumulated sound effects library directly into the timeline, supporting natural language search and Find Similar sound effect lookup. The extension requires Pro Tools 2026.11 or later, and after installation can be used from the Extensions menu without opening the Soundly desktop app.
Google released the Gemini Omni Flash 1.2 video model on November 5, 2026, and it is now available on Recraft Studio. Compared to 1.1, the single generation duration has increased from 10 seconds to 15 seconds, consistency and lip-sync are clearer, and Gemini API prices have dropped by about 30%.
Why it matters: The new version increases the single generation duration from 10 seconds to 15 seconds and improves cross-segment consistency and lip-sync, making it suitable for short ad scenarios that need a complete mini-story.
Adobe announces an upgrade to its creative suite, extending Adobe Creative Agent to Creative Cloud and Firefly, and providing local MCP integration for Premiere, After Effects, and more.
MiniMax H3 Image receives day-0 native support in ComfyUI, with a single checkpoint supporting both text-to-image and reference image editing. The model is CFG-distilled, requires 6 sampling steps and no negative prompt, natively outputs resolutions such as 4096×4096 (1:1), and supports multi-reference image editing and prompts up to 4096 tokens.
Why it matters: MiniMax H3 Image receives day-0 native support in ComfyUI, with a single model handling both text-to-image and reference image editing, 6-step sampling, and native high-resolution output, enabling creators to evaluate whether to replace their existing image generation workflows.
WellSaid launches Producer, embedding professional voice production into agent workflows such as Claude, Copilot, and ChatGPT via an MCP server, and natively generates 48kHz audio. The platform supports agent style guides based on over 500,000 Oxford Languages transcriptions, open design, and digital human replication, with the license retained with the final audio.
JJYB_AI_VideoAutoCut v3.4.6 fixes export pipeline consistency issues: when multi-format export fails, it returns failure; for partial success, it lists results and errors separately; native drafts respect explicit output directory; and fixes issues such as global voiceover across clips repeating from the beginning. Full pytest suite: 110 tests passed, including 21 new multi-format export regression tests and 15 draft export API/frontend state regression tests. Real-media end-to-end smoke tests verify speed-change source intervals, continuous voiceover cursor, and subtitle material consistent with meta duration.
Adobe releases Photoshop 28.0, adding three AI features: Relight, Remove related elements, and Generative Recompose.
The author ported the ReCaVSR one-step streaming video super-resolution model based on Wan2.2 TI2V-5B to a native ComfyUI node, supporting 4×/2× super-resolution, executing one diffusion step every 8 frames.
MiniMax H3 Image provides a prompt-writing skill, downloadable from the official GitHub. This skill is similar to the existing skill for video models, helping to structure prompts and improve generation consistency. Download address: https://github.com/MiniMax-AI/MiniMax-H3-Image/tree/main/skills/h3-image-prompt-writing。
Adobe MAX 2026 opens, and Premiere receives updates including the official availability of Color Mode and the Generative Media Tool. The Generative Media Tool brings video, sound effects, soundscapes, and music generation directly into the timeline, and supports Firefly as well as partner models from Google, Kling, Runway, Luma, and others.
Why it matters: Within the Premiere timeline, third-party models such as Kling and Runway can be directly invoked to generate missing footage, further integrating video editing and AI generation workflows.
Adobe updates Firefly and Premiere mobile at MAX 2026, bringing new features for video creation, image editing, and mobile editing. Firefly Video can turn travel footage into vlogs, Creative Agent can reside in the background to handle repetitive tasks; Premiere mobile adds UMG and WMG music partnerships and arrives on Android.
Adobe announced at MAX 2026 that local MCP integration will arrive in Creative Cloud tools in early 2027, allowing third-party AI agents to work directly in native project files of Premiere, Photoshop, Illustrator, and After Effects.
Adobe is adding Creative Agent to Premiere Pro, which can automatically apply effects, track masked subjects, and adjust color and exposure across an entire sequence. The new Generative Media Tool can generate video, sound effects, and music directly within the timeline, and supports integration with third-party models such as Kling; Generative Remove can remove objects from the frame, and Color Mode provides a color grading workspace.
Why it matters: Adobe is integrating third-party models such as Kling into the Premiere timeline, allowing editors to directly generate assets, but the copyright boundaries of third-party models are more ambiguous than Firefly.
comfyui-workflow-templates v0.11.105 has been released to PyPI. It can be installed via `pip install comfyui-workflow-templates==0.11.105`. This update includes adjustments to the MiniMax H3 Image workflow (PR #1409).
Adobe previews local MCP integration for Creative Cloud at MAX, allowing third-party AI agents such as OpenAI, Google, and Anthropic to directly edit Photoshop, Illustrator, Premiere, and After Effects project files.
Why it matters: Adobe previews local MCP integration for Creative Cloud, enabling third-party AI agents to directly edit project files and boosting automation in creative workflows.
Avid releases Ready to Play, partnering with Google Cloud to integrate Gemini and Veo-powered AI video generation and analysis directly into Media Composer.
open-ai-canvas v1.8.4 released, updates cover creation homepage, channel management, video request logs, short drama delivery, canvas and Agent, and login protocol. Database Schema upgraded from 92 to 98, adding MUSE2API and SeedanceMINI1 asynchronous video protocols, each provider supports up to 100 recommended models.
AI Canvas v0.11.1 unifies the basic parameters for images and videos of the same model, and adapts to Wanxiang video duration and reference materials. Wanxiang H3 only supports 15, 20, 25, 30 seconds, default 15 seconds; C2/C3 only support 5, 10, 15 seconds, C3 prohibits audio reference. Also improved Muse upload and send confirmation, added material node 'Copy File Path', old projects are not automatically migrated.
MOSS-Image ecosystem releases multiple updates, including new Control LoRA such as OpenPose, Lineart, Depth-map, and 461 new characters. It also launched the Character Sheet Designer workflow and Mineral Ink, Risograph style LoRA. The Fizgig trainer reduces LoRA training time to about 20 minutes.
agnes-video-generator released v8.0.0, refactored the interface and added quick retry, one-click simple video, and external material collaboration modes. The external material mode can import and concatenate clips generated by external tools such as Jimeng while preserving the original audio tracks; the quick retry mode can retry 1-120 times per minute when the video queue is full. The one-click simple video defaults to using agnes-video-2.5-flash to generate 5-second 16:9 720P clips.
AIArtEngine releases version v8.9.1, with changelog covering v8.9.0 to v8.9.1.
Starting November 9, 2026, Apple will allow Apple Creator Studio subscribers to purchase a $49.99 AI usage pack in Logic Pro for AI features beyond the monthly quota, available in eight countries including France.
AIArtEngine v8.9.0 released; the changelog covers changes from v8.8.0 to v8.9.0.
Many thanks to @shiyi_c98 for helping us bring Pika H3 Flex inference to AMD GPU 💜 We look forward to collaborating with more researchers in the community. If you are an AMD developer, we would love for you to contribute to this branch! Check out her work: https://github.com/Pika-Labs/Pika-H3-Flex/tree/contrib/amd-mi350x
The author released a 2-step LoRA for LTX-2.5 Stage 1, reducing Stage 1 from 8 steps to 2 steps, and achieving end-to-end accelerated rendering with SoL Refiner and PrunaAI VAE decoder.
The ComfyUI port of the ReCaVSR video super-resolution model has been released. int8 quantization allows 1440p 4x super-resolution to be completed on 8GB VRAM. The model is based on Wan2.2 TI2V-5B and comes with built-in FlashDecoder streaming decoding, so long clips only increase time and disk usage. In testing, a rainy street scene from 960×480 to 3840×1920 took about 36 seconds on an RTX 4090, with the int8 file at 6.27GB and output differing from bf16 by about 48dB.
Why it matters: This ComfyUI port uses int8 quantization to compress 1440p 4x super-resolution to run on 8GB VRAM, allowing low-VRAM users to assess deployment based on this.
CreativeField is a free and open-source ComfyUI interface that integrates image generation, video, 3D, and photo editing into a single workspace. It supports generating images with FLUX or Qwen, generating videos with MiniMax H3 or HyperFlow, converting images to 3D with Hunyuan3D or TRELLIS, and includes a built-in FotoField photo editor.
The author released a custom ComfyUI long video node that breaks Viggle Remake's single-run 345-frame limit through overlapping windows and stitching. The default window is 294 frames (12.25 seconds at 24 fps); subsequent passes start from the last 17 frames of the previous pass and remain fixed, with audio split and stitched at the same intervals. The node does not detect shot cuts; seam continuity still requires manual inspection, and the source video must be 24 fps.
Runway launches Sightline feature, which can create 360-degree panoramas of any location, with the environment remaining consistent when framing from any angle; the exported panorama can be used for 3D scene environments, and is now available in App Studio.
CRLFF Digital releases audio plugin Katharsis X, designed to repair vibrato, codec hiss, phase high frequencies, and missing octaves in AI-generated music and stems. The plugin features built-in neural network guidance, can analyze and repair per frequency band and per time instant, offers 8 repair modules and simple/advanced modes, supports Windows and macOS, with an introductory price of $99 until December 31, 2026.
ComfyUI's October update brings multiple efficiency improvements including branch nodes, LoRA stacking, automatic template downloads, multi-output zip export, EXR/HDR/ProRes support, link organization, and in-widget undo/redo, all included in v0.44. Branch Node supports up to 32 inputs, LoRA stacking node supports up to 20 LoRAs, and template downloads are limited to official templates.
Why it matters: ComfyUI's October update brings multiple efficiency improvements such as branch nodes, LoRA stacking, and link organization, allowing creators to optimize workflows, reduce repetitive operations, and improve creative efficiency.
Otoy has begun rolling out the public beta of OctaneRender 2027.1. The new version adds a real-time neural rendering mode that delivers near-noise-free interactive rendering at low sample counts. The beta is available alongside the stable 2026 release, and adds AI Light 2.0, USD export, and MCP Server support.
agnes-video-generator (free-short-video) v7.5.1 released, fixing the image generation task details page and progress display, and adding multi-architecture Docker images.
InvokeAI 7.0.0 enters beta 1, its largest version update. It adds projects, a single workbench, and video panel support for MiniMax H3, LTX-2.5, and Wan 2.2. Official warning: v7 database migration is irreversible; back up before upgrading.
Why it matters: InvokeAI 7.0.0 enters beta, adding projects, a video panel, and multi-model support, giving creators a basis to evaluate the upgrade.
Fizgig 9.2 was released on November 8. The time to train MOSS-Image LoRA on an RTX 5090 (94 images, batch 1, Muon, 70 epochs) was shortened from 49 minutes to about 20 minutes, and memory usage dropped from 21.7 GB to 7.1 GB.
GLITCH Studio is recruiting 1-2 technical beta testers to test its ComfyUI-based local AI filmmaking Windows application. The application supports workflows such as WAN, MiniMax H3, Krea 2, LTX 2.5, and FLUX, and has completed clean machine installation tests for FLUX image editing and LTX 2.5 video rendering.
MiniMax H3 Novel Pipeline is preparing to release version 2.5.1, refactoring the ComfyUI node structure and improving the character profile editing process. Updates include H3 reference budget (9 images, 3 audio files), 4-15 second scene limits, spatial continuity JSON Schema validation, and OpenAI/LM Studio backend configuration. Node logic has been separated into src/nodes/ and src/core/ for easier future maintenance and expansion.
Developer BittiAI built slopfab, a native command-line tool that lets coding agents directly generate media files, supporting models such as MiniMax H3, Krea 2, MOSS, Fibo 2.0, YuE2, and MiniMax Music3, covering video, image, and music generation. The tool can run locally on Windows/Linux, supports CUDA/Vulkan, requires no cloud or account, and can be executed remotely via SSH.