The workflow shown in the image above utilizes the open-source text-to-image Flux model, known for its realistic rendering, and the latest model Gen-3 from the leading AI video generation company Runway.
In the past, this workflow required creators to manually transfer materials between different platforms, consuming a lot of time. Especially as the number of materials increased and the size of the creative team expanded, the difficulty of material management also increased.
Now, you just need to open Kaiber, click into the new feature "Superstudio" to create a canvas, and you can easily set up your preferred workflow. In addition to Flux and Runway, this tool also integrates popular AI video models such as Luma, Keling AI, and Conch AI.
Entering the main page of Kaiber's new feature "Superstudio" (the relevant links have been organized at the end of the article), the official site provides two workflow examples. One is an image style synthesis combining reference images and prompts, and the other is a text-to-video workflow that introduces the Luma video generation model.
You can click on the left side of the interface to add a new blank canvas and build your own workflow.
After creating a blank canvas, it's normal to feel a bit lost at first. Simply familiarizing yourself with the three core operations will allow you to get started quickly.
First, call the model or workflow.
By clicking on the "Flow Menu" in the top menu bar, you can find all the models and workflows integrated into the product. There are popular mainstream video models available, as well as workflows for image or video style enhancement, re-drawing, and other functionalities.
Select your desired model or workflow and drag it into the canvas.
Second, in the material storage area on the right, you can find all the materials you have uploaded or generated.
If there are too many materials, you can also organize them into a collection (Collections) by project or material type.
Compared to previous tools where material storage was relatively independent, this design is very intuitive. You can easily filter the materials you need to reuse by dragging them, without having to open local folders or search through the tool's material library.
Finally, for specific generation operations, it works just like using other AI video tools—by inputting text prompts or uploading images and videos, you can train the AI to do the work.
Some complex workflows can also be built and reused more easily than before, greatly improving creative efficiency.