AI video generator with sound: the clip comes with its audio
Three of the video models on Visuall make sound together with the picture: Gemini Omni Flash (people speak the line you put in quotes), Seedance 2.0 (ambience and effects, the best physics) and MiniMax H3 Max. Veo 3.1 and Kling are silent. Below, clips with their sound, made on the Visuall canvas, and the honest list of which model does what.
Clips with their sound, and the prompts
Each clip came from the prompt under it with its sound made by the model: speech on Gemini Omni Flash, the pour and the waves on Seedance 2.0. They start muted; turn the sound on with the controls.
Gemini Omni Flash · 6 s · speech
“A chef in a busy kitchen looks at the camera and says: "Three ingredients, five minutes. Watch." Pans sizzle behind him.”
Gemini Omni Flash · 6 s · speech
“A young traveller films a selfie in front of the Colosseum and says: "You won't believe how big this is." The camera stays on her face.”
Seedance 2.0 · 5 s · sound
“Cold brew coffee is poured over ice in glossy close-up, condensation on the glass, studio light, the ice cracks and the coffee swirls”
Seedance 2.0 · 5 s · sound
“A golden retriever runs along a beach at sunrise, low tracking shot, waves splashing, warm light”
More in the gallery.
How it works
- 1
Pick a model with sound
Gemini Omni Flash for speech, Seedance 2.0 for ambience and physics, MiniMax H3 Max for value.
- 2
Write the sound into the prompt
Put spoken lines in quotes. Name the sounds you want: "pans sizzle", "waves", "a crowd cheering".
- 3
Leave Sound on
Every video box has a Sound switch. On, the model makes audio with the clip; off, the clip is silent.
- 4
Stitch and keep the sound
Stitch Videos keeps each clip's audio, and Captions draws over the picture without touching it.
AI video models and prices
All on the same credits. Credit packs start at $5 for 100 credits, and new accounts get 35 free. Compare every model →
| Model | Price |
|---|---|
| Veo 3.1 LiteGoogle720p · 4, 6 or 8 s · silent | 8 credits / 4s clip |
| Kling 2.6 ProKuaishouImage to video · 5 or 10 s · silent | 20 credits / 5s clip |
| MiniMax H3 MaxMiniMax768p · 5–15 s | 27 credits / 5s clip |
| Veo 3.1 FastGoogle720p · 4, 6 or 8 s · silent | 22 credits / 4s clip |
| Kling 3.0Kuaishou3–15 s · multi-shot · silent | 27 credits / 5s clip |
| Gemini Omni FlashGoogle720p · 3–10 s · speech & sound | 33 credits / 5s clip |
| Seedance 2.0ByteDance720p · 4–15 s · sound | 52 credits / 5s clip |
| Veo 3.1Google720p · 4, 6 or 8 s · silent | 52 credits / 4s clip |
Which AI video models have sound
- Gemini Omni Flash: speech, ambience and effects. Write the line in quotes and the person says it.
- Seedance 2.0: sound made with the picture, strongest on physical sounds, up to 15 seconds.
- MiniMax H3 Max: sound with the clip, the default model of Animate Image.
- Veo 3.1, Veo 3.1 Fast, Veo 3.1 Lite, Kling 3.0 and Kling 2.6 Pro: silent. Cheaper per clip, and you add music or voice yourself.
When silent is the better choice
A film cut to a song you will add yourself is better silent: a model cannot sing your song, and invented audio under it is worse than none. Turn Sound off in the box, or use Veo 3.1, and add the music in your editor.
Frequently asked questions
Which AI video generator has sound?
On Visuall, Gemini Omni Flash, Seedance 2.0 and MiniMax H3 Max make sound with the video. Gemini Omni Flash also makes people speak the line you put in quotes. Veo 3.1 and Kling clips are silent.
Can the AI make the person talk?
Yes, with Gemini Omni Flash. Write the words in quotes in the prompt, keep a line to about 20 words for an 8-second clip, and say where the camera stays.
Does the sound survive stitching and captions?
Yes. Stitch Videos keeps every clip's audio in order, and Captions and Logo draw on the picture only.
Is it free to try?
New accounts get 35 free credits with no card, enough for a 5-second Gemini Omni Flash clip with speech. You see the price before you press Run, and a failed clip is never charged.
Try it free
35 free credits when you sign up. No credit card required.