More control. More models. A new connection to 3D.
The latest Node Composer release brings important improvements to the way you build, organise and expand visual AI workflows. From automatic workflow organisation to new image models and a new Blender integration, this update makes it easier to move from individual generations to connected creative processes.
Auto Align
Keeping a complex node workflow organised shouldn’t get in the way of creating. Auto Align automatically arranges nodes into a cleaner structure, making connections easier to read and workflows easier to navigate.
Whether you’re working with a small experiment or a larger production workflow, you can now spend less time organising nodes and more time composing your process.
New image models
We’ve expanded the image-generation options available inside the Image node with new models and variants.
Midjourney
Midjourney is now available as an image-generation option, giving you another powerful visual model to incorporate into your workflows. It runs through WaveSpeed, returns four images per run, and takes a style reference to guide the look of the result.
GPT Image 2.5
This release also introduces two new GPT Image 2.5 variants:
- Sunburst — the slower of the two. The extra time goes into fine detail and into text that comes out readable. The one for the final run.
- Flare — the quicker one. A little less fine detail, back much sooner. The one for drafts and variations.
Same price, same resolutions, same sixteen reference slots: the only difference is the time spent on each image. Both are available on WaveSpeed and Runware.
Having different models available as nodes means you can experiment with different approaches, compare results and combine generations within the same visual workflow.
The model is no longer the workflow.
It’s one component of it.
New connector: Blender
The new Blender connector brings Blender scenes directly into Node Composer.
Instead of treating 3D production and AI generation as separate processes, you can now use a Blender scene as part of a connected creative workflow. The node talks to the copy of Blender running on your machine, reads the scene you have open and renders it right there on the canvas — flat-shaded, in seconds, at no cost.
A typical workflow might look like:
Blender scene → 3D render → AI generation → Refinement → Final output
Your Blender scene provides the 3D foundation — geometry, composition, camera and production setup — while the AI nodes downstream explore different visual directions, materials, environments and treatments. Wire the render into a Video node and the model works from your camera move instead of guessing one.
This opens up new possibilities for:
Architectural visualisation · Interior design · Product design · 3D art · Creative production
And importantly, the Blender connector is not another generative model.
It’s a connection between the tools you already use and the AI models inside your workflow.
Bug fixes and improvements
This release also includes a number of minor bug fixes and stability improvements across the application.
Small changes, but important for keeping workflows running smoothly.
Build. Connect. Compose.
With new models, new connectors and better workflow organisation, Node Composer 3.0 continues to evolve from an AI generation tool into a visual system for composing creative production.
New nodes. New connections. A stronger way to compose.