Text to video clips with sound from a prompt

Text to video is Deepscene's open clip tool. The prompt is the film. Reference images are optional, up to nine, and they go out with the words you wrote. Seedance 2.5 is the default model. Kling 3.0 is the other choice. A new clip starts at 5 seconds, 16:9, 720p, with audio on.

How to use it

1. Open Text to video.
2. Write what happens, in order, including sound if you want speech, effects, or a bed.
3. Add reference images only when a face, product, or place must match something you already have.
4. Pick the model, length, ratio, and resolution, then generate.
5. After the clip lands, a note edits that file.

Kling 3.0 stays at 720p and holds at most 15 seconds. Seedance 2.5 can hold a single generation up to about 30 seconds. The file you download is the clip the model returned.

Where to use it

Use Text to video when you want a clip that is not a short, a reel, or an ad: a product beat, a mood piece, a background loop, or a test of a scene. Prefer Image to video when a picture you already have is what should move. Prefer Hypervisuals when the start frame and the end frame are both locked. For a cast, a script, and several cuts, use Short and film.

What you pay

Seedance 2.5 is 5 credits a second at 720p and 2 credits a second at 480p. A 5 second clip at the 720p default is 25 credits. Kling 3.0 with audio is 4 credits a second, and silent Kling is 3. Starter is $15 a month for 150 credits, enough for about six of those 5 second 720p clips, or about 75 seconds if you drop Seedance to 480p. Creator is $39 for 420 credits. Studio is $99 for 1,100. You can cancel any time. There is no setup fee. The full table is on pricing.

Leave audio on when the clip should carry speech, effects, or a music bed. Turn audio off only when you want the picture alone.

Questions

Is the soundtrack replaced after the model finishes?

No. The clip you save is the file the model returned, sound included.

Do I have to upload a picture?

No. References are optional here. If the picture is the starting point, use Image to video instead.