Text / Image to VideoRegenerate
A latte, from pour to first sip
Text to video AIWrite a scene. Get a video clip.
Review results, manage assets and access generation history in AI Studio.
Text to video AI Write a scene. Get a video clip.
Text to video on ImgAI generates a short clip from a written description alone. You describe the subject, action, setting, camera movement and sound; the model builds every frame from those words, with no photo or footage to start from. Pick a video model, aspect ratio, duration and resolution in the generator above, and the credit cost shows before you submit.
These clips started from text only
No image was uploaded for any clip below. Open one to load its prompt, model and settings into the generator and use it as a starting point.
Every clip below is a real job run here. Tap one and the prompt and stills load — run it yourself.
Turn an idea into a clip
What to make with text to video
Text is the fastest start when the shot does not exist yet: nothing to photograph, nothing to upload.
Ad concepts before a shoot
Describe the product, the setting and one action, then compare versions before you book a studio. A prompt such as “A perfume bottle on black stone, one fine spray crosses a beam of light” is enough to test the idea.
Storyboards and establishing shots
Turn script lines into short shots to check pacing, framing and mood. Generate each shot separately, then put them in order in a video editor.
Vertical clips for social feeds
Choose 9:16 for Reels, Shorts and TikTok, or 1:1 for feeds. Put the hook at the start of the prompt so it appears in the opening seconds.
Scenes you cannot film
Weather, historical settings, impossible camera paths, or a product in a place it could never be. Describe them the same way you would describe a real location.
AI video models you can use
Compare models by input type, target clip length and output quality. The cards show current options and default costs; check the settings again when switching models.
Video creation
Give your idea some movement
Consider the clip length and resolution, then choose a model. Each price uses the settings shown below.
Veo 3.1 Fast
Newly added8-second clips with sound at the best value
- Up to 3 reference images
- 720P
- 1080P
- 4K
- Up to 8 seconds
Cost at default settings
200 credits
16:9 · 720P · 8s
Veo 3.1 Lite
Newly addedThe fewest credits — test an idea from text first
- 720P
- 1080P
- 4K
- Up to 8 seconds
Cost at default settings
100 credits
16:9 · 720P · 8s
Veo 3.1 Quality
Newly addedFinal cuts that need the best picture
- Up to 2 reference images
- 720P
- 1080P
- 4K
- Up to 8 seconds
Cost at default settings
1430 credits
16:9 · 720P · 8s
MiniMax H3
Newly addedMiniMax
Turn the clip in your head into a first version
- Up to 5 reference images
- 768P
- 2K
- Up to 15 seconds
Cost at default settings
408 credits
16:9 · 768P · 5s
Seedance 2.5
Newly addedByteDance
Leave more room for a longer clip
- Up to 30 reference images
- 480P
- 720P
- 1080P
- Up to 30 seconds
Cost at default settings
687 credits
Auto · 480P · 5s
Seedance 2.0
ByteDance
Try another model for a different take on your clip
- Up to 9 reference images
- 480P
- 720P
- 1080P
- Up to 30 seconds
Cost at default settings
472 credits
Auto · 480P · 5s
Wan 3.0
Newly addedWan
Describe the scene, action and camera movement
- Up to 10 reference images
- 480P
- 720P
- 1080P
- Up to 30 seconds
Cost at default settings
290 credits
Auto · 480P · 5s
Model cards show capability ranges, not a promise that every setting can be combined. Frame inputs and reference media may change the available options. Check the generator for the settings that apply to your clip.
How to turn text into a video
- 01
Write the scene
Describe the subject, one main action, the setting, camera movement, light and sound. For clips of 10 seconds or more, split the action into timed beats.
- 02
Choose the model and format
Pick a video model and a fixed aspect ratio such as 16:9 or 9:16, then set duration and resolution. Options depend on the model, and the credit cost updates as you change them.
- 03
Generate, review, revise
Track the clip in AI Studio. Check motion, framing and sound, then download it, or change one part of the prompt and generate again.
How to write a text to video prompt
Write in shot order
Subject, then action, then setting, then camera, then light and sound. Someone reading the first sentence should be able to picture the first frame.
Split longer clips into timed beats
For a longer clip, describe the order of events, such as “[0–3s] she looks left, then right; [3–6s] she reaches for the glass.” Timing cues guide the model but may not be followed precisely.
Spell out speech and sound
Put dialogue or voice-over in quotation marks and say who speaks and in what voice. List sound effects and music in a separate sentence. After generation, check the words and the lip sync.
Say what must stay out
Name unwanted elements such as on-screen text, extra characters or camera cuts. Keep the list focused and check the result, since exclusions may not be followed fully.
How much does an AI video cost?
Video generation requires an account and credits. The cost depends on the model, output duration, resolution and inputs such as reference video. Try a shorter clip to check the shot before choosing a longer or higher-resolution version. The generator shows the cost before submission; credits are available through subscriptions or one-time packs.
Text to video FAQ
No. Leave the upload slot empty and the clip is generated from your prompt alone. If you later need a fixed opening shot, a product's exact look or the same character across clips, add an image and switch to image to video.
The limit depends on the model, from 3,000 to 20,000 characters; the generator enforces the limit of the selected model. Length is not the goal: a few clear sentences usually beat a long list of adjectives. For longer clips, timed beats add detail without contradicting each other.
Models with audio generation produce sound together with the picture, and some let you turn it off. Write dialogue in quotation marks, name the speaker and describe sound effects separately. Spoken words and lip sync can drift, so review them before you publish.
The usual causes are too many actions in one clip, conflicting camera instructions, or several events competing for the same moment. Reduce the clip to one main action, give the camera one movement, and change one thing per attempt. A completed clip that differs from your intent is not a failed task; tasks that fail return their credits.
Not reliably with text alone: each generation interprets the description again, so faces, clothes and products vary between clips. To keep a subject consistent, generate or upload an image of it and use image to video or reference images.
No. Video generation requires an account and credits; signup bonuses and other credit offers can be used under their rules. The cost depends on model, duration and resolution and is shown before you submit. A short, low-resolution test costs the least.