ComfyUI

Bringing #TencentCloud #HY3D natively into ComfyUI's node-based ecosystem means creators no longer have to juggle disconnected tools — text-to-3D, image-to-3D, and sketch-to-3D workflows are now just nodes away. The open-source community meets enterprise-grade 3D infrastructure. We can't wait to see what creators build with this. 🚀

Try your own training on AI Toolkit!

We just released Topaz Starlight Precise 2.5 as a Partner Node in ComfyUI — and it's the most meaningful upgrade to their upscaling model yet. If you're working with video upscaling pipelines, here's why this matters: 🔹 Sharper fine details — especially visible in 720p → 4K workflows 🔹 Fewer artifacts — cleaner results with reduced hallucination compared to SLP-2 🔹 More convincing output — the model fills in realistic detail that makes upscaled footage look native, not stretched 🔹 Same price — no cost increase for better quality 🔹 Drop-in replacement — just select `Starlight Precise 2.5` from the model dropdown. No workflow changes needed. Getting started is simple: 1. Update ComfyUI or open Comfy Cloud 2. Find the Topaz upscaling node (or load the Video Upscale template) 3. Select Starlight Precise 2.5 and run your workflow Already using Starlight Precise? Swap to v2.5 in the dropdown — same cost, better results, no friction. 🔗 Try on Comfy Cloud https://lnkd.in/gK6pG5SS

The Grok Imagine updates keep accelerating. Reference-to-Video + Video Extend just landed! Live in ComfyUI on Day 0. The generation speed alone makes it worth trying. Reference-to-Video: add up to 7 reference images to generate consistent characters and scenes Video Extend: seamlessly extend any video with smooth, natural motion Grok Reference-to-Video Workflow https://lnkd.in/gUpFurBb Grok Video Extend Workflow https://lnkd.in/gGJk5RXR Read more about the new Grok updates on our blog https://lnkd.in/gm5EmRjA

🎞️ LTX 2.3: The AI Film Studio - Deep Dive 🎤 Hosts: Purz & Fill ⏲️ March 25th – 3pm PST / 6pm EST 📍 Live on YouTube, X, and Twitch We're doing a full deep dive on LTX 2.3 and testing every modality — text & image to video, first/last frame, and audio-driven generation. https://lnkd.in/gssyEGT3

Turn any image + audio into a talking video with LTX 2.3 🎤 Just drop in a photo and a voice clip… and it will generate a video, animating your image to match the audio with really solid lip-sync. It’s now easier than ever with the new ComfyUI app. Try it today https://www.comfy.org/

We’re hiring engineers, designers, and operators at ComfyUI. One thing we care deeply about is how people think about AI tools. Many companies are building systems meant to automate creative work but our view has always been a little different. We’re trying to give creators more control over the process itself. You can see that philosophy directly in the product. ComfyUI doesn’t try to hide the system. The workflow is right there, and if you want to modify it, you can. It’s the kind of environment that tends to attract builders who enjoy experimenting with systems rather than treating them as black boxes. If that way of thinking resonates with you, let’s talk comfy.org/careers

Style transfer tooling has come a long way — from IP-Adapters and complex image-to-image workflows to image-edit models, dedicated style transfer models, and UI friendly LoRA trainers. Each approach can suit different use cases and with so many methods now available, it helps to have a clear guide. We put together a Style Transfer Handbook to map out what's available today and how to think about choosing between different techniques. Check out the blog for more details! https://lnkd.in/gJvQMAk7