How to create a video from text with Flow AI Video

To create a video from text, open Text to Video AI, describe one scene, choose a video model and supported settings, then check credits and generate. A reference image is optional for this starting workflow.

What should the prompt include?

Name the subject, an action, the setting, and a camera direction. Keep the first take to one visible moment instead of an entire story. Details such as lighting and background help define what you will inspect afterward.

A ceramic mug on a wooden kitchen table at sunrise. Steam rises slowly while the camera moves gently closer. Keep the mug steady and the background softly out of focus.

The subject is the mug, the action is rising steam, and the camera movement is a slow approach. If the take is too busy, remove an instruction before adding more detail.

Which settings do I choose?

Pick the model before choosing duration, aspect ratio, and resolution. The menu shows the settings supported by that option. Choose a portrait or landscape format for the place you intend to publish; avoid assuming every model offers the same controls.

Sign in and review the current credit estimate. A longer duration or different model can change the cost. If credits are insufficient, the dialog lets you compare plans without starting the request.

What happens after I click Generate?

A successful submission takes you to Create, where your prompt and selected settings accompany the task. Review progress, open the completed video, and download the take you want. View more history below the landing-page input is another route to Create.

How do I improve the next take?

Watch the full clip against the original brief. Check the subject, motion, framing, and unwanted objects. Change one instruction, such as “keep the camera locked,” then compare the next result. You can also move to Image to Video AI when a photo explains the composition better than words.

Create a video from your text.