Bug: Low Cost Narrated Video Generator

[Townsend]

This workflow creates a narrated video, without ever using the expensive Generate Video block. It goes like this.

1- User Input: describes a scene as a narration.
2- Generate Image: based scene description.
3- Text to Speech: generates audio narration of scene.
4- Analyze Video (Audio): retrieves length of audio narration in seconds.
5- Generate Static Video from Image: using previously generated image
6- Mix Audio into Video: Combines image with audio narration into video.

I’m submitted this as a Community Agent. It’s a cool workflow that any user can use to generate a narrated video for just the cost of generating the image. (about 0.04)

Point 1: The Analyze Video block can also be used to analyze audio. In this case retrieving the length of the audio track in seconds.

Point 2: In order for this narrated video to match the length of the audio track, the length of the audio track in seconds must be entered in the Generate Static Video from Image block.

Bug: And even though the Duration field of the Generate Static Video block accepts a variable, shown by a handlebar prompt, using a variable here always crashes the workflow.

This is the workflow that I’ve submitted as a Community Agent. The Duration value is set to a literal 5 seconds, so the video being generated will be clipped short. Just replace that ‘5’ with {{seconds}} and you will be able to recreate the crash.

https://app.mindstudio.ai/agents/bug-low-cost-narrated-video-generator-ff8abc77/run

[Alex_MindStudio]

Hi @Townsend,

Thanks for the post!

The Generate Static Video From Image block doesn’t support variables in the Duration field yet. I’ve logged a feature request with our engineers and will follow up as soon as I have an update.

[Alex_MindStudio]

Hi @Townsend,

You can now reference variables in the duration field of the Generate Static Video block.

Thanks for taking the time to write up the post!

[Townsend]

Thanks! Now we can make NARRATED VIDEOS with static images for like 5¢ a minute. I made a few minor modifications, replacing the error description with developer instructions.

If you’re interested… the above workflow link should should be enabled for Sharing and Remixing.

Note: OpenAI Text to Speech. (Not Eleven Labs which is expensive.)