How to make YouTube Shorts using Lewis

Keytalk AI
Lewis AI
Published in
4 min readFeb 9, 2024

Discover how to use Lewis to create engaging and viral YouTube Shorts that captivate your audience.

Step 1. Conceptualize Your Short

Choose an idea that aligns with your audience’s interests and your channel’s theme. Ensure it has the potential for high engagement and shares.

For example, creator KakuDrop uses photorealistic AI-generated images to create an interesting short music video:

We’ll take a similar approach but make a compelling short AI film instead of a music video.

Step 2. Script Your Short with Lewis

You can get everything you need to create a Shorts or Reels video from Lewis.

Input your concept into Lewis to generate the world and scenes from your story. I chose this vampire adventure story I had generated earlier for the purpose of this tutorial.

Refer to this article on how to create a story using Lewis if you’re new to this series.

Now back to the video:

You can find all your projects in your dashboard, which you can access from anywhere on Lewis by clicking your profile module.👇

From your dashboard, go to ‘projects’ and sort through projects to find the one you’re after.

You can publish the project if you wish to share the link with others or keep it private in your dashboard — it’s your choice!). 👇

Publishing your project allows you to create a sharable link for easier collaboration and promotion.

If you want to add voice-over to your movie, make sure you generate a script while you’re generating scenes. 👇

When you generate your scene, you can choose to generate a script for the scene.

Don’t worry if you’ve already published your story without a script because you can always update your published story by hitting the ‘publish this project’ button once again.

Step 3. Export your images to a video editor

Select the images you want to use in the video from your project page and hit the download button. Your image will be saved to your local folder in PNG format.

There’s a little download button for every image on your project page.

You can use the images as-is, or head to Runway ml to turn your image into 4-second video clips. It’s as simple as creating a free account and dropping your image into the image box. Adding a description is an option.

Hit the ‘start with image’ button and simply drop your image into the box.

Next, we’re moving on to video editing. I am using Canva for this tutorial, but you can use any video editor with a basic animation feature. The great thing with Canva is that you can access cinematic templates like these. 👇

Follow these steps:

  • Create a mobile video format
  • Select a template (optional)
  • Upload the scene images from Lewis
  • Add animation effect to each image
  • Add a background music

We already have a great video, but I really recommend that you follow the next step to generate a narrative voice for the video.

Step 4. Add narrative voice

Return to your project page and jump to the ‘scenes’ section. You will find the script module at the bottom of each scene box if you have already generated a script. 👇

If you have generated scenes but not scripts, you can always use the ‘edit’ feature at the top of the scenes section and then scroll down to the scene you want to add a script to. 👇

‘Regenerate’ will replace your current scenes with a new set of scenes, so make sure you ‘edit’.

Select three to four sentences you want to use for the narration and copy them to your clipboard manually. 👇

Manually drag the text you want to use for the video and copy. Feel free to edit.

To create realistic voices, go to Elevelabs. Using the free plan, you can generate up to 10,000 free characters worth of speech generation. 👇

Under settings, you can select a voice and listen to the sample speech. Hit ‘generate’ when ready.

Download the file and drop it in your video editor. Sync it to flow well with your video. Adjust the volume of the background music, if necessary, to ensure you can hear the narration clearly.

Keep the narration to three to four sentences max if your video is less than a minute.

The final result should look like this:

--

--

Keytalk AI
Lewis AI

Keytalk AI is the most advanced prompt engineering company