Text to video AI
Describe the subject, action, setting, and camera movement. Generate a clip in ChatCut, then shape it into a finished edit.
Trusted by 400,000+ creators
How it works
How to use text to video AI
- 01
Describe one scene
Write your shot request on this page, open ChatCut, and sign in free. Include the subject, action, setting, and camera movement; no image upload is needed for text generation.
- 02
Choose Video Gen settings
Switch to Video Gen and select the model name. Open the Video settings slider, choose text generation under Input, then set Aspect ratio, Resolution, and Duration.
- 03
Generate, review, and edit
Check the credit estimate beside Send and select Send message. Follow progress in My Assets, play the finished clip, and download it or drag it to the timeline.
Why ChatCut
Why use text to video AI in ChatCut
-
A prompt is enough
Start with a written scene and no uploaded file. The AI video generator takes your subject, action, setting, camera direction, and visual style as the brief.
-
Choose Seedance or Kling
Select Seedance 2.5 or Kling 4.0 in Video Gen. Each model exposes its supported inputs, duration, aspect ratio, and resolution before you send.
-
Set the shape before generating
Choose a wide frame for a landscape or a vertical frame for a social clip. Available shapes depend on the selected model, so recheck Video settings when switching.
-
Review each take’s estimate
The estimate beside Send updates with the model and settings. Each successful request makes one clip; another take is a separate generation with its own credit charge.
-
Keep the clip in your project
The finished result appears in My Assets. Download the clip directly when its card is ready, or drag it onto a video track to build a sequence.
-
Finish around the generated shot
Combine the take with your footage and add a voiceover or background music. Manual timeline edits and export do not consume credits; generation and Agent messages do.
Tips
Write a clearer text to video AI prompt
Describe a visible action, its timing, and its ending before adding style details.
-
Choose one visible action
Name the subject, action, setting, light, and one camera move so the prompt describes a coherent shot.
-
Fit actions to the duration
Divide longer scenes into short time ranges and name the ending so the action fits the selected duration.
-
Revise one thing per take
Watch the whole result, then change the failed movement while keeping successful lighting and setting details to focus the next attempt.
Use cases
Text to video AI ideas to try
-
Food details for a recipe edit
Add a close-up that gives a recipe sequence texture and a moment to breathe. This Seedance 2.5 shot uses one action, honey dripping onto a honeycomb, with warm backlight.
-
An invented wildlife insert
Create a wildlife insert for a fictional story or a narrated explainer. For this Seedance 2.5 lion close-up, the prompt specifies wind through the mane, golden light, and a single subject.
-
A short dialogue scene
The watchmaker and granddaughter scene starts from a written prompt with timed beats and a line of dialogue. Keep the exchange short enough that the reaction has room to register.
-
A reveal for an opening sequence
Open a story with a close-up, then reveal the setting. This Seedance 2.5 shot pulls back from fireworks reflected in a woman’s eyes to a harbor lit by the display.
Keep editing
Build the next shot on the same timeline
Arrange generated shots beside your footage, trim the edges, and leave room for narration. Make cover artwork that matches the finished story.
- Conversational editing
- Text-based editing
- Auto-captions
- Motion graphics
- AI voiceover
- Generate clips
- Noise removal
- Multi-track timeline
-
You Can Now Edit Videos with Grok Bot
Edit with Grok Bot -
How to Make 3× the Content With AI (No Editing Experience)
Turn long videos into shorts -
I Turned One Product Photo Into a Full Ad With ChatGPT
Product ads skill -
How to Create Insane Motion Graphics with Astra 6
AI motion graphics
Skills
Hundreds of skills for any video
100+ Skills. No experience needed. Start in one click.
FAQ
Text to video AI FAQ
Is generation free?
Video generation requires Pro and uses credits, as do built-in Agent messages. Uploads, manual editing, transcription, and export do not consume credits, though plan limits apply; see the credits policy.
Do I need a photo or an existing video?
No. Text generation needs no upload. When a person or product must remain recognizable, supply an image through a supported reference mode instead.
How long can one generated clip be?
The duration options depend on the model. Kling 3.0 supports 3 to 15 seconds, and Kling 4.0 supports 3 to 30 seconds. For a longer story, generate separate shots and arrange them on the timeline.
Can one request make a complete long video?
Each successful generation returns one video asset. Break a long script into shots, generate each moment, and assemble the clips and narration afterward.
Can I choose the file format?
Seedance 2.5 offers MP4 or MOV in Video settings. Other models have their own output settings. Downloading an asset and exporting an edited timeline are separate actions.
Can the generated clip include sound?
Kling 3.0 and Kling 4.0 generate native audio. Describe dialogue, sound effects, or music in the prompt; those models do not take audio references. Play the result with sound before deciding whether to keep it.
Can I generate another version of the same scene?
Yes. Submit the prompt again for another take. Each is a separate job, asset, and charge, so change only the details you want to test.
Can I download and edit the result?
Yes. Wait for completion, then use Download in My Assets. For a sequence, drag the clip to the timeline, add other media, and use Export.
Related
More text to video AI and editing tools
All AI apps
Audio & voice
Captions, text & translation
AI video & images
All free tools
Convert & compress
Edit video
Audio & voice





