ComfyUI

Stable Audio 3.0 is Day-0 in ComfyUI—open weights on fully licensed data, from quick effects and short tracks to pieces up to about six minutes, in the graph you already use. Small SFX, Small, and Medium cover different jobs. Set the duration in seconds in the template. Outputs fall under the Stability AI Community License for commercial work. For many creators, that means one workflow for lo-fi or synthwave ideas, instrument loops, footsteps, and ambiance, or short one-shots—not three separate tools. For more details: https://lnkd.in/gjSDsCSW What would you try first—SFX, a loop, or something longer? #GenerativeAI #CreativeTech #AIWorkflows #ComfyUI

We just raised $30M at a $500M valuation and we're hiring the PM who'll help shape what we build. ComfyUI started as the tool users reached for when nothing else gave them enough control. Now we're building Comfy Cloud, the Hub, and App mode on top of that core. The person we need can hold both in their head. The power user who lives in the node graph, and the professional creator who hasn’t found ComfyUI yet. As a product manager here, you'll develop agentic user experiences for our 4 million creators, build vibe-coded prototypes, and dig into bugs in code yourself. We're a small team in San Francisco and you'll work directly with the founders on everything you ship. Roles are open now: https://lnkd.in/dx5dKPzM

📽️ LTX 2.3 IC LoRAs - LipDub & Video Utilities 🎤 Hosts: Purz & Rob ⏲️ May 20th – 3pm PST / 6pm EST 📍 Live on YouTube, X, and Twitch Today we're diving into LTX 2.3 IC LoRAs to check out LipDub and a bunch of other Video Utility LoRAs trained by the community! https://lnkd.in/gKDuZGwd

Using Seedance 2.0, a simple performance shot can become a cinematic scene no matter where it was shot. Workflow link below. https://lnkd.in/gfVFaAUZ

Excited to announce that Anthropic's Claude is now available as a partner node in ComfyUI. If you build image and video pipelines, this unlocks a new layer of intelligence inside your workflows. No custom code. No bolted-on API calls. Just drop Claude directly into your workflow and put frontier reasoning exactly where you need it. → Prompt enhancement — turn a one-line idea into a richly detailed prompt before it hits your image or video model A few things to highlight: → Image understanding — pass an image in, get back an accurate description, critique, or targeted edit suggestions → Workflow logic — generate captions, alt text, scene breakdowns, or branching decisions that downstream nodes can act on Powered by frontier models like Opus 4.7. Outputs a clean string, so it slots into anything that takes text. Pair it with your favorite checkpoints and you've got a workflow that can actually think about what it's generating. Try it inside your workflow today: https://lnkd.in/gnHWt6we

From months of work at ILM to a single compositor in ComfyUI — that's not just a workflow change. That's a fundamental shift in who gets to do this work. Doug Hogan and the ActionVFX team built this course around that reality. Advanced AI Workflows for VFX Artists is exactly what professional adoption looks like. Traditional VFX pipelines, AI tools, one unified workflow inside ComfyUI. This is a great example of why we built the platform.

We believe everyone deserves to design exactly what they dream, own their process, and never compromise because of their software. Our team behind Comfy is small, intense, and building what is now becoming the industry standard of the future. We're looking for people who want to build tools that make others more capable, not just more efficient. People who care more about output than title, want real agency in a flat org, and would rather ship and learn than plan and wait. If that's you, we're building something worth your time. Our values: → Ship it — Work hard for what makes it into the world. Ideas don't substitute for output. → Share it — Every decision and its reasoning gets shared with the team. → Open-source it — If it can be open, it should be open. That's how we build an ecosystem that moves faster than any company could imagine. → Respect the craft — We build for people who prefer precision over shortcuts. The learning curve is high, but the payoff is worth it. If this is how you already think, we want to hear from you. View open roles: https://comfy.org/careers

A breakdown of a ComfyUI workflow that combines Seedance 2.0 with an LLM prompt setup designed for cinematic motion shots like this. This example recreates the viral floating hot sauce effect with a Mexican hot sauce twist. Workflow link below. https://lnkd.in/gUHX4xtj

Native support for the best open source models. Three open-sourced models now have native ComfyUI support: → Gemma 4 by Google DeepMind - A natively multimodal open-weights model supporting text, image, audio, and video input, with a configurable thinking mode for step-by-step reasoning before output. → VOID by Netflix —-Video inpainting that goes beyond erasing pixels. Remove a subject and VOID also removes the shadows, reflections, and physical interactions it caused in the scene. → BiRefNet - High-resolution background and object segmentation from CAAI AIR 2024. One of the most widely adopted segmentation models in open-source workflows today. Workflows and blog in the link below 👇 https://lnkd.in/gFmW3B_h