Define your
subject.
Describe who or what appears, its position and the details that matter. Without an image, words establish the visual starting point. Be specific about appearance, but avoid competing lists of styles.
AI Text to Video Generator
Describe a scene. Create a short video without a reference image.
Start with words.
Build a scene.
Turn a written description into a short AI video. No reference image is required: define the subject, give it a clear action and decide how the camera should see it. Think in individual shots rather than a whole screenplay.
Start with a promptDefine your
Choose one
main action.
Give the subject a readable action, direction and pace. A pirate raising a lantern is a more focused starting idea than an entire adventure with several locations, characters and cuts.
Decide what
the camera does.
Separate camera movement from subject movement. Choose a fixed view, a gentle push or a following shot. These are prompt instructions, not dedicated camera controls, and may need refinement.
Everyday actions.
Clear interactions.
For a coffee-making idea, describe one action and the objects involved: a barista pours milk into a cup while the camera stays close. Check hands, the cup and the flow of liquid throughout the result. If the action is unclear, simplify the interaction before adding more detail.
Moving subjects.
Separate directions.
For a chase scene, say who moves, where they travel and whether the camera follows. If both the subject and camera seem to move unpredictably, first try a fixed camera and one clear direction of travel. Change one instruction at a time so you can judge its effect.
Fast actions.
Manage the complexity.
For a football idea, focus on one player and one action rather than a full match. Describe the framing and the ball’s intended movement, then inspect contact, body shape and continuity. Extra resolution does not guarantee that an incorrect action or distorted detail will be fixed.
From a rough idea.
To a clear prompt.
“A cinematic train in the snow” suggests a mood, but leaves the action and camera unclear. Build it into a focused shot with the five details below. This is a teaching example, not the original prompt or generation record for any displayed asset.
Start with a promptEstablish the starting scene.
Name the subject, location and time of day: “A passenger train crosses a stone viaduct in snowy mountains at night.” Include the details that define the shot before decorative adjectives. Text gives the model a direction, not a guarantee of a particular identity or exact appearance.
Make the action explicit.
Add direction and pace: “The train travels from left to right at a steady speed.” This answers what changes over time. Start with one main action; several arrivals, turns, collisions and scene changes can make a short shot harder to assess.
Specify the viewpoint.
Choose framing and camera behavior: “The camera stays fixed in a wide view, keeping the bridge in frame.” This removes ambiguity about whether the viewer moves with the train. When motion is confusing, try a fixed view before introducing tracking or an orbit.
Add supporting atmosphere.
Add light and secondary movement: “Warm light glows through the windows while snow falls gently.” Keep these details subordinate to the train’s action. Choose a coherent visual style instead of combining contradictory instructions such as bright midday light and a dark night scene.
Combine and refine.
Complete teaching prompt: “A passenger train travels from left to right across a stone viaduct in snowy mountains at night. The camera stays fixed in a wide view, keeping the bridge in frame. Warm light glows through the windows while snow falls gently. Keep the shot continuous, without cuts.” If the result changes scenes, remove competing events and restate the single shot. Change one detail per attempt; prompt wording cannot guarantee continuity.
Three scenes.
Three camera choices.
These still-image concepts illustrate ways to plan a shot; they are not playable video results. Use them to distinguish a moving subject from a moving camera. The suggested directions are teaching examples, not promises of exact reproduction.
A fixed camera.
A passing train.
Describe a train crossing the bridge from left to right while a wide camera view stays still. This separates the train’s movement from the framing. Check that the bridge, mountains and direction of travel remain stable across the clip.
A following camera.
A flying falcon.
Describe a falcon flying low over the waves, with the camera following beside it at a steady distance. Give the flight a clear pace and direction. If the bird drifts out of frame, simplify the shot and clarify its intended position.
A gentle push.
A quiet shopfront.
Describe a slow approach toward the flower shop while the building stays still and light reflects on the wet street. Keep background activity minimal. This makes camera movement—not a busy sequence of events—the main change in the shot.
How to turn text into video.
Start a draft here, then continue in the workspace to sign in, confirm your prompt and settings, and generate. On mobile, choose your output settings in the workspace.
Write a focused shot
Describe a subject, one main action, a setting and a camera direction. No upload is required. If a particular photo or product appearance needs to guide the result, consider AI Image to Video instead.
Confirm settings and cost
Choose 3–12 seconds, a supported aspect ratio and a 480p, 720p or 1080p output setting in the workspace. Check the credit cost before submitting. For an overview of input choices and video specifications, see the AI Video Generator.
Review, refine and save
Watch the full result for action, framing and continuity. If it misses your idea, revise one prompt detail before trying again. Download the completed MP4 within 72 hours of completion and keep a local copy.
Frequently
asked
questions
Can I generate a video using only text?
Yes. A written prompt is enough; a reference image is optional, not required. Start your draft on this page and continue in the workspace to submit. The shared video generator also accepts reference images if you choose to add them. Pure text is useful for exploring a scene from scratch, but cannot guarantee an exact product design or a specific person’s appearance.
What should I include in a text to video prompt?
Describe the subject, its action and direction, the environment, camera behavior and lighting. Prioritize one clear shot over a list of unrelated events. The prompt field allows up to 4,096 characters; a longer prompt is not automatically better. Camera directions are written instructions rather than separate camera settings. Review the result to see which instructions need clarification.
Can I turn a full script into a finished video?
This workflow generates short clips from scene descriptions, not an automatically edited film from a long script or article. Break a larger story into individual shot ideas. There are no dedicated timeline, subtitle, narration, clip-extension or video-editing controls in this workflow, and consistency between separate clips is not guaranteed. Generated audio, dialogue and lip synchronization are not promised here.
What settings and free allowance are available?
Select 3–12 seconds in whole-second increments, with 1:1, 16:9, 9:16, 3:2 or 2:3 for text-only input and no automatic ratio option. Output settings are 480p, 720p and 1080p; Grok Imagine 1.5 generates at the selected setting without a separate upscaling step. The default is 5 seconds, 480p and 16:9. New accounts receive 50 credits valid for 30 days: if unused, these cover one default 5-second 480p video for 40 credits, with 10 credits remaining. At 480p, 3, 5 and 12 seconds cost 20, 40 and 90 credits. Cost changes with duration and output setting; check the workspace before generating. The allowance is not daily or unlimited.
How can I improve a result, and when should I download it?
If an action is unclear, simplify it and specify direction or pace. If the camera is distracting, try a fixed view; if scenes change unexpectedly, remove extra locations and events. Review body shapes and object interactions as well as the overall style. Change one instruction at a time when generating again. Completed videos are available as MP4 downloads in the workspace for 72 hours after completion. Save a local copy before expiry; generation time is not guaranteed.
Can I use the videos commercially, and are they private?
Free and paid generated outputs may be used commercially under our Terms of Service. You are responsible for rights and permissions relevant to your inputs and intended use; this does not guarantee exclusive copyright. Prompts, uploads and outputs are not used to train AI models under our platform policy, and uploads and outputs are private by default. See the Privacy Policy for details.