1. Describe the first moment
Sign in, switch the box to Image, write the opening frame and send. Pick the image model in the picker; the still's price is beside it.
Text to Video
01 · Words to a frame
First the still. You see it before any clip exists.
Describe the opening moment and pick an image model: 8 are in the picker, each with its credits beside it. The still comes back in the frame you chose, and you can draw it again or edit it until it is the picture you meant.
- A still costs a fraction of a clip, so the wrong guess is a cheap one.
- Ten frames, from 9:16 to 21:9.
02 · Frame to a shot
Then the clip, starting on that frame.
Turn the still into a shot and write what happens next. The studio writes the scene as timed shots held to your still, and each shot renders as its own clip on the video model you pick. The clip opens on the frame you approved.
- The studio does not render a clip from words alone. A scene needs at least one picture on it.
- The still can also be animated directly, with a camera move or a template effect.
03 · The price first
Two prices, each shown before it is spent.
The still's price is beside the model that draws it. The clip's price is on the scene's Render card, for the model, length and frame you picked. Neither runs until you press its button.
How to go from text to video
2. Turn the still into a shot
Under the still, choose Turn into a shot and write what happens. The studio writes the scene as timed shots held to that still.
3. Approve the render
Open the scene's Render card, pick the video model, the length and the frame, read the price and approve. The clip lands in your Library.
From a Sentence to a Shot
- 01
Write it plainly
One line is enough: who or what, where, and what happens. The studio writes the full scene prompt, the camera and the sound from it.
- 02
A first frame from your words
The opening still is drawn from the description alone, on the image model and in the frame you choose.
- 03
Change one thing
Edit the still with a sentence, or draw a variation, before it becomes the start of a clip.
- 04
Timed shots
What happens next is written as timed windows, 4 to 30 seconds in all, each one shot with its own clip.
- 05
The clip opens on your still
Image to video, from the frame you approved: the model continues a picture you have already seen.
- 06
The model is your choice
LTX, Wan, Kling, Seedance or Veo for the clip, picked on the Render card with its price.
- 07
Talk it through first
Not sure what the scene is yet? The assistant works through the idea with you and writes the prompt when you ask it to.
- 08
Sound, written in
Every scene prompt ends in its own sound: the footstep, the room. LTX 2.5, Seedance 2.5 and Veo 3.1 render it with the picture.
- 09
Kept in your Library
The still and the clip both land in your Library with the prompt and the model that made them.
9 Video Generator Models
LTX 2.3
Fast, cheap, 1080p and up. The workhorse. 87 credits for a 6-second 1080p clip, on every plan.
Wan 3.0
Open-weight realism up to 1080p. 120 credits for a 5-second 720p clip, on every plan.
Veo 3.1
Google's top model, with sound. Premium only. 576 credits for a 6-second 720p clip, on the Studio plan.
9 Models
Every video model the studio renders on, LTX 2.5 Fast, LTX 2.5 Pro, Kling 3 Turbo Pro, Seedance 2.0 Fast, Seedance 2.0, Seedance 2.5 included, with the price on the card before you approve.
See every model
FAQs about Text to Video
Start with a sentence
Write it. See the frame. Then the shot.
The still comes first, so you know what the clip starts on.