# We’ve been testing a bunch of AI video and lip-sync tools. Here’s a 28-second music video we made — retained record

> Retained reddit record from reddit:generativeAI.

Canonical: https://highsignal.app/data/records/56ba90c74baf52bf

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:generativeAI)  published 2026-09-30 12:42 ingested 2026-10-01 02:43

[original source ↗](https://www.reddit.com/r/generativeAI/comments/1wu3xj4/weve_been_testing_a_bunch_of_ai_video_and_lipsync/)

## retained text



```
We’ve been trying a bunch of video generation and lip-sync tools to see how they work together in a music video. Here’s one of the results: a 28-second clip for “Satellites in the Rain.”



For context, we’re the team behind MUVO. Both the song and the video were generated with the app. The person in the clip is a synthetic demo character, and we made this example for onboarding.



Here’s what’s going on under the hood:



• Gemini 3.5 Flash helps with lyrics and musical direction, then Mureka 9 generates the song.

• Our own prompt-enhancement layer is a big part of the work. We’ve put R&D into turning a simple idea into detailed, model-specific prompts and coordinating the different steps.

• For the video, MUVO maps out timed shots and Gemini helps flesh out the direction. This clip has four shots.

• GPT Image 2 creates a shared visual reference and storyboard images for each shot to help keep the look and character consistent.

• Two shots use Kling Avatar v2 Standard for lip-sync performance; the other two use Seedance 1.5 Pro. The video models run through fal, and the text/image calls go through OpenRouter.

• FFmpeg trims and stitches the clips together, removes their generated audio, and puts the song excerpt over the final edit.



Curious what you think: which part looks convincing, and where does the illusion fall apart? If you’ve been experimenting with AI music videos too, what’s worked for you?



App details if you want to take a look: [https://muvoai.com](https://muvoai.com)
```



document fetched 2026-10-01 00:42 · canonical https://reddit.com/r/generativeAI/comments/1wu3xj4/weve_been_testing_a_bunch_of_ai_video_and_lipsync
