Google DeepMind · Video generation
Turn a short scene brief into video with native sound. Use Veo 3.1 on Astorie to develop a cinematic moment, explore dialogue or bring a visual starting point into motion, then review the take in your canvas.
Create cinematic motion from text, generate native sound and explore light across detailed materials.
Describe the subject, action, lighting and camera movement in one brief. Veo 3.1 generates a moving scene from text, making it useful for exploring staging and shot rhythm before a shoot. Keep the action focused and inspect the result for physical and visual continuity.
Veo 3.1 can generate dialogue, ambience and sound effects with the video. Place sound directions beside the action they accompany, from a brief spoken line to a tool touching a surface. Review the complete take for wording, pronunciation and synchronization.
Describe how light moves across a surface and how the material should respond. Veo 3.1 can render cinematic scenes with detailed textures, making it a starting point for glass reflections, metal highlights or fabric studies. Keep the setup simple, then inspect reflections, edges and surface detail throughout the take.
Bring your first Veo 3.1 shot to life in three steps.
Use Open in canvas to open a project, then select Veo 3.1 in a video node. The model picker also lists Veo 3.1 Extend separately, so check the selected name before continuing.
Connect a text node describing the subject, action, camera and sound to the video node. Check the quality mode, aspect ratio, resolution, duration and credit estimate before starting generation. The video node shows the run in progress.
Play the completed result in the canvas. Review movement, details, dialogue and ambience, then revise the prompt for another take or download the result you want to keep.
Develop dialogue beats, tactile craft films and location studies with picture and sound planned together.
Explore a short performance built around a pause, a glance and one spoken line. Keep the speaker visible and the staging simple so the take helps the team discuss delivery and timing. Confirm the words and lip movement before choosing a direction.
Develop a close view of a maker at work, with a single tool action and a distinct sound. Use the clip to explore tactile details and the balance between the action and the surrounding room tone. Inspect contact points and tool geometry before editing.
Build a short establishing shot around a place changing over a few seconds: a curtain lifting, leaves moving or light crossing a wall. Combine a restrained camera move with location sound to test the atmosphere of a sequence.
Google's model capabilities and the controls available in Astorie are listed separately.
| Provider and version | Google DeepMind Veo 3.1; Astorie model ID google-veo3.1. This page covers the base model.Source ↗ · 2026-11-09 |
|---|---|
| Astorie input modes | Text to video, Image to video and Reference to video are available in the Veo 3.1 model picker.Source ↗ · 2026-11-09 |
| Native audio | Google documents native dialogue, ambience and synchronized sound effects. Review the generated audio before use.Source ↗ · 2026-11-09 |
| Google output specifications | Google documents 4, 6 or 8 seconds, 16:9 or 9:16, 720p, 1080p or 4K, and 24 fps. Reference-image generation uses 8 seconds; availability varies by variant and mode.Source ↗ · 2026-11-09 |
| Astorie output configuration | The page Generator uses the model defaults. Check the output and audio settings offered by your selected canvas mode, and review the estimate before generating.Source ↗ · 2026-11-09 |
| Video continuation | Astorie lists google-veo3.1-extend-video separately from the base google-veo3.1 model. Use the matching video-input workflow for extension.Source ↗ · 2026-11-09 |
Start with text for an invented scene or use images to give the shot a visual anchor.
| Dimension | Text-led scene | Image-guided scene |
|---|---|---|
| Starting materialSource ↗ · 2026-11-09 | A written scene brief; let Veo invent the setting and subject. | A starting image in Image to video, or reference material in Reference to video. |
| Direction to writeSource ↗ · 2026-11-09 | Describe the subject, action, camera and sound together. | Explain what the image should guide and what should move; add the sound you want. |
| Review focusSource ↗ · 2026-11-09 | Check that the staging, timing and sound match your brief. | Check reference details, edge stability and motion as well as the sound. |
Plan a focused take with enough time to review its movement, details and sound.
使用最先进的 AI 模型开始创作。免费计划无需信用卡。
用 3 天体验 Pro 的全部功能
Try another video model when you want to develop a different interpretation of the brief.
Plan the script, shot list and edit that turn individual clips into a short film.
Inputs, sound and the path from configuration to a generated take.
Veo 3.1 is a Google DeepMind video generation model with native audio. On Astorie, select Veo 3.1 to create a scene from a text brief or use the image inputs supported by the chosen mode.
No. Open in canvas creates a project with the selected configuration; generation and credit use happen in the canvas. If you are signed out, sign in or register to continue with the configuration.
Yes. The model supports native dialogue, ambience and sound effects. Describe the sound explicitly and check the available audio settings in the canvas. Listen to the exported take before approving the words or timing.
Yes. Astorie lists Image to video and Reference to video for Veo 3.1. Choose the matching mode and supply its required input. Reference guidance does not guarantee that faces, logos or fine details remain unchanged.
Astorie lists Veo 3.1 Extend as a separate model. This page's default is the base Veo 3.1 model, google-veo3.1. Select a workflow that accepts an existing video when continuation is the task.
Check the estimate for your configuration and the price shown in the canvas before you generate. An estimate is not a record of credits charged; current plans are shown in the plans section.
Treat them as starting directions to adapt to your scene. Keep one clear action, describe the camera and sound, and check that the brief fits the selected duration. They are creative suggestions rather than generation logs for the adjacent examples.
Choose Veo 3.1, describe the action and sound, then take your configuration into the Astorie canvas.
