Video Generation
Generate video in Studio from text or images, with start/end frame control.
Studio generates video from a text prompt, from a starting image, or both. You can control the first and last frame, choose duration and resolution, and pick the model that best fits the look you're after.
Video generation always runs as a background job, so you can leave the page and return to find it finished.
Choosing a model#
Studio offers several video models with different strengths — cinematic motion, character consistency, realism, precise frame control, or fast drafts. Some also support extras like reference media or built-in audio. The picker explains what each model is best at, and the settings around it adapt to the model you pick.
| Model | Best for | Duration | Resolution | Frame & reference control |
|---|---|---|---|---|
| Seedance 2.5 | Long shots, up to 30s | 4–30s | 480p–720p | Start + end frame, or image/video/audio references |
| Seedance 2 | Cinematic motion | 4–15s | 480p–1080p | Start + end frame, or image/video/audio references |
| Kling 3.0 | Character consistency | 3–15s | 1080p | Start + end frame |
| Veo 3.1 | Realism | 4, 6, or 8s | 720p–4K | Start frame |
| Veo 3.1 Lite | Fast, cheaper drafts | 4, 6, or 8s | 720p–4K | Start frame |
| Wan 2.7 | Precise frame control | 2–15s | 720p–1080p | Start + end frame |
| Hailuo 2.3 Pro | Expressive character action | 6s (fixed) | 1080p (fixed) | Start frame |
| Happy Horse 1.1 | Native audio and lip-sync | 3–15s | 720p–1080p | Start frame, plus up to 9 image references |
| Grok Imagine 1.5 | Quick creative experiments | 1–15s | 480p–720p | Start frame |
| Gemini Omni Flash | Prompt-based video edits | 3–10s | 720p (fixed) | Start frame, or attach one video to edit |
The lineup evolves over time, so the picker in the app is always the source of truth — new models appear there before they appear here.
Settings#
- Prompt — Describe the motion, subject, and style. Optional if you supply a start and/or end frame.
- Aspect ratio — Square, 16:9, 9:16, 4:3, 3:4, and more depending on the model.
- Duration — 1–30 seconds depending on the model (most top out around 15; some offer fixed durations only).
- Resolution — 480p, 720p, 1080p, or 4K depending on the model.
- Start frame — Provide an image to animate (image-to-video), or let the model generate one from your prompt.
- End frame — Some models let you set a final frame to control where the shot lands.
- Reference media — Some models accept reference images (and, in some cases, video or audio) to guide character and style.
Image-to-video#
To animate an existing image:
Choose a start frame
Upload an image, or click Use frame on any image in Studio (including a still pulled from another video) to set it as the start frame. The video inherits the image's aspect ratio where possible.
Describe the motion
Add a prompt describing how the scene should move. With models that support end frames, you can also set where the shot should finish.
Generate
Submit and let the job run in the background. Watch the progress indicator and queue position; the finished clip lands in your session and library.
Audio and lip-sync#
Some video models generate native audio and lip-sync, so dialogue and sound are baked into the clip. Otherwise the video is silent and you can pair it with music or sound effects from Studio.
Using your videos#
Finished videos are saved to your Asset Library as MP4 and can be downloaded or dropped into a Canvas design.