Turn a written scene into a video
Describe your subject, setting and action. Choose a video model and format, then bring the scene into motion without a starting image.
In a misty ancient forest, an East Asian warrior in dark blue robes stands in a shallow river as a giant tiger emerges between the trees. The tiger circles him, then charges through the water. The warrior dodges with his wooden staff. Track the action in one continuous cinematic shot. Cold blue light, drifting fog, realistic fur and splashing water. River ambience and tiger growls. No dialogue, music or text.
Read the full scene prompt
Create a hyper-realistic, dark fantasy action sequence set in an ancient, mist-covered forest with towering trees, moss-covered rocks, cascading waterfalls, and a shallow river. A rugged East Asian warrior with long wet black hair, weathered skin, traditional dark blue battle robes, leather armor, and a wooden staff struggles through the rushing water. A gigantic supernatural tiger emerges from the fog, its muscular body covered in dark brown and black striped fur, glowing fiery red eyes, enormous fangs, and an aggressive, predatory expression. The warrior grips his staff tightly as the beast slowly circles him, growling and preparing to attack. Scene progression: Begin with a wide establishing shot of the warrior kneeling in the river. Move into close-ups of his determined face, trembling hands gripping the staff, and the tiger's glowing eyes. The tiger suddenly charges through the water, creating a massive splash. The warrior dodges the attack and turns to face the beast as it roars ferociously. Capture dynamic tracking shots, slow-motion water droplets, dramatic close-ups, and wide shots emphasizing the creature's immense size. Visual style: Photorealistic 8K cinematic quality, ancient Chinese dark fantasy, realistic wet fur and skin textures, volumetric fog, atmospheric lighting, cold blue-gray color grading, natural water physics, detailed facial expressions, dramatic contrast, shallow depth of field, anamorphic lens, 35mm film aesthetic, epic creature design, intense suspense, Hollywood-level VFX. Camera: Smooth cinematic tracking, low-angle shots, handheld action movements, slow-motion impact shots, extreme close-ups, wide environmental shots, seamless transitions. Audio: Deep tiger growls, powerful roars, rushing river water, splashing water, heavy breathing, staff movements, forest ambience, and tense cinematic sound design. No dialogue and no background music. Negative prompt: Cartoon, anime, low resolution, blurry visuals, plastic skin, unrealistic fur, deformed anatomy, extra limbs, duplicated characters, floating objects, unnatural water movement, flickering, inconsistent character appearance, distorted faces, text, subtitles, logos, watermarks.
Give your idea a first take
Create before you have a photo
Explore a character moment, product setting or opening shot directly from a written idea.
Direct the action and camera
Describe what the viewer sees and what moves. Keep the action focused on the time available in your clip.
Refine one detail at a time
Reuse a take, adjust the framing or movement and compare the results before downloading.
Turn a description into a video
- Step 1
Write your scene
Describe the subject, setting, action and camera movement. Give the model the details that matter to your idea.
- Step 2
Choose your video settings
Select a model, aspect ratio and duration. Review the credit quote, then generate when you are ready.
- Step 3
Watch, download, repeat
Preview your video beside the panel. Download the result or reuse its inputs to refine your next take.
Text-to-video FAQ
No. This page offers text-to-video generation. Describe the scene and choose the available model and video settings.
Choose from the durations shown for your selected model. The available length and resolution depend on the model and other settings.
Audio support depends on the selected model. If audio controls are available, choose the setting and describe the sound you want in your prompt.
Your video appears beside the tool when it is ready. Sign in to browse Video History, open a task’s details or download a finished clip. Choose Reuse to restore its inputs for another take.
Generation uses credits. The panel shows the cost for your selected model and settings before you submit. Visit Pricing for current plans and credit packs.
Creations are public by default. Eligible paid accounts can turn off Public before generating. Check the setting before submitting.
A confirmed failed generation returns its reserved credits. An interrupted connection does not mean a task has failed. Check the task status before submitting again.
Files and results are kept for 10 days for free accounts and 180 days for eligible paid accounts, measured from upload or creation. Download what you want to keep before it expires.
