The generator offers one configured generation workflow in each input mode. There is no extra model decision to make in this composer. For model-specific workflows, visit the video model detail page or browse all models; their controls and prices may differ.
Supported settings by input mode| Setting | Text to Video | Image to Video |
|---|
| Input | Required written prompt | Required start image and motion prompt; optional end image |
|---|
| Length | 4–15 seconds, whole seconds | 4–15 seconds, whole seconds |
|---|
| Resolution | 768P or 2K | 768P or 2K |
|---|
| Framing | 21:9, 16:9, 4:3, 1:1, 3:4, 9:16 | Follows the source image; no separate ratio control |
|---|
| Upload | No image needed | JPG, PNG, WebP; up to 20 MB per image |
|---|
| Prompt limit | 7,000 characters | 7,000 characters |
|---|
Choose a clear source with a readable subject and enough space for the intended motion. Very small text, complex hands, fast camera moves, and precise product geometry need careful review. We do not promise exact identity preservation, lip synchronization, or physically accurate motion. There is no separate audio switch in this workflow; inspect any sound in the finished clip. Generated output is not a substitute for factual or rights review.
If you already have footage and only need more resolution, use the video upscaler. Generating a scene and enlarging an existing clip solve different problems. For longer edits, assemble reviewed clips in your preferred editor rather than expecting one request to deliver an entire production.