The AI Video Generator has just been released as a research preview. This article aims to answer any questions you might have about how to use it, what it is capable of, and what our plans are going forward.

The AI Video Generator is powered by an agentic system called Cora. Cora acts as an orchestrator, understanding your intent and then designing a storyboard for you.

Layouts and scripts are dynamically generated, making the system highly flexible and capable of working with images included in your sources.

You can also chat with Cora to make adjustments to your storyboard before you open it as a draft, for even more fine-grained tweaking in the Colossyan Video Editor.

 💡 This release is a research preview, and this feature is in active development.

How to create a new video with the AI Video Generator

To access the AI Video Generator (Research Preview), select the AI Video Generator Button on the home page.

You’ll be given the option to try out the new AI video generator, but if you want to use our previous version of document-to-video or script-to-video, you can still access these with the buttons below.

Once you select "Try it now!", you’ll be taken into the Cora chat window.

Here you can start by attaching documents. Currently, we support:

PPT
PDF
TXT
DOCX

We have a file limit of 5GB.

You can upload multiple files at once if you want to.

You can also provide a prompt. Our guidance here is to be as explicit as possible. Describe the video you want in as much detail as possible. Highlight key themes you want the video to focus on, or key messages you want delivered.

Finally, you can select a brand kit. If you don’t have one, we recommend creating one, as it will ensure that your fonts, typography, logo and colours are used in the generation.

Once you are ready, select the arrow on the bottom right.

Once you have done so, you’ll see that any documents you have provided are being processed.

This can take a bit of time, especially if the documents contain many pages or slides. The process will create summaries of the documents and extract images from your sources, which can be used in video generation.

You’ll also notice that Cora will ask you several contextually sensitive questions to help ensure it understands what it needs to create for you.

After this, you’ll be taken into the storyboard view. Here you can see previews of the scenes on the left, the scripts per scene on the right, and on the far right, your agentic chat window.

💡 Preview generation can take a bit of time, so please bare this in mind.

Once the preview generation is complete, there are a few ways you can tweak the storyboard before further manual editing.

Select a part of the script, and edit it manually.
Select part of the script, and add it to the chat to target edit with ai.
Select the AI button on the top of any script box, and add that entire scene's script to the chat for targeted editing with AI.
Select any object on the preview to do a spot edit with AI. You can also quickly delete the selected object.
In the case of an image or video included in a preview, you can also replace it with stock media or videos, or images extracted from your source files.
You can make global changes by describing the changes you want made in the Cora chat.

Beyond the various editing options, you can also add new scenes with the button at the bottom of the storyboard editor, as well as drag and drop scenes to reorder them.

Once you are ready to do some manual tweaking and generate your draft, select the ‘open as draft’ button on the top right corner. This will open the draft as a new tab in the Colossyan Video Editor.

How to find a storyboard when you return to Colossyan

If you have finished up for the day and come back later, you can find all the storyboards you’ve created via the plans area on the home page.

You’ll always be able to see your most recent three plans here.

If you select ‘all videos’, you’ll also be able to see all the video plans you’ve created.

Current limitations

While the AI Video Generator represents a large improvement in functionality compared with our existing document-to-video and script-to-video workflows, there are some limitations you should be aware of currently.

It does not currently support creating interactive videos. This is something we are working on.
It cannot read images. This means that if you ask Cora to create a video with the design in a style that is represented in an image or PPT, it will not be able to do so. This is something we will address soon.