AI video answer
How do I write a text-to-video API prompt?
Write one filmable shot in this order: subject, visible action, setting, camera, light and style, timing, then dialogue and sound. Keep duration and output settings in API fields.
Start with subject and physical action
Name who or what the viewer sees, then describe one action that can finish inside the clip. Use concrete movement such as turns, lifts, walks, pours, opens, or looks instead of abstract mood words alone.
State the setting and the objects that must remain visible.
Direct one camera move
Choose framing, angle, camera path, speed, and landing frame. A slow push toward a medium close-up is easier to stage than several competing moves in one short shot.
Add light, color, texture, and production style after the physical scene is clear.
Separate prompt from settings
Put duration, aspect ratio, resolution, audio toggle, seed, model, callback, and budget in structured fields. Put spoken words, sound effects, and ambience in clear prompt sentences when the route supports audio.
This split makes the request easier to validate, compare, and rebuild on another model.
Related guide
Get the full context.
Text to video API with model routing, audio, and cost controls
Compare text-to-video API routes, prompt fields, audio support, pricing, jobs, and output controls across Wan, MiniMax, Kling, Seedance, and Veo.
Read the text to video API guideCurrent prices
Compare each video route.
| Model | Compared rate | Cre8vid price | 5 sec | 10 sec | Savings |
|---|---|---|---|---|---|
| MiniMax H3MiniMax | $0.065 | $0.029 | $0.145 | $0.29 | 55% less |
| Wan 3.0Alibaba | $0.050 | $0.025 | $0.125 | $0.25 | 50% less |
| Kling 3.0Kuaishou | $0.084 | $0.039 | $0.195 | $0.39 | 54% less |
| Seedance 2.5ByteDance | $0.144 | $0.069 | $0.35 | $0.69 | 52% less |
| Veo 3.1 Fast + audioGoogle | $0.100 | $0.049 | $0.24 | $0.49 | 51% less |
Source notes