Subject
A ceramic robot florist
Text → shot → job → video
Turn one filmable prompt into a priced, trackable video job. Pin a model or route by audio, duration, resolution, availability, and maximum cost.
{
"mode": "text_to_video",
"prompt": "A ceramic robot florist...",
"duration_seconds": 10,
"aspect_ratio": "16:9",
"resolution": "720p",
"audio": "required",
"route": "cheapest_compatible",
"max_cost_usd": 0.50
}Prompt anatomy
Keep the prose about the shot. Keep duration, ratio, resolution, model, audio, callback, and budget in typed API fields.
A ceramic robot florist
ties a red ribbon around one bouquet
inside a rain-lit shop at blue hour
slow push from wide shot to medium close-up
soft window light, paper rustle, one bell chime

Text-to-video route prices
Compare the same duration and output settings. Every rerun is another generated attempt.
| Route | From | 10 seconds | Use |
|---|---|---|---|
| Wan 3 | $0.025/s | $0.25 | Lowest listed start |
| MiniMax H3 | $0.029/s | $0.29 | Low-cost second family |
| Kling 3 | $0.039/s | $0.39 | Controlled motion route |
| Veo 3.1 Fast | $0.049/s | $0.49 | Listed native-audio route |
| Seedance 2.5 | $0.069/s | $0.69 | Higher-cost review route |
Reject unsupported settings.
Return route and total.
Create one idempotent job.
Return output and final cost.
Text-to-video questions
A text-to-video API lets software send a written scene and output settings, start an asynchronous generation job, track its status, and receive a video result without using a visual editor.
Read the full answer →The listed Cre8vid routes include Wan 3, MiniMax H3, Kling 3, Seedance 2.5, and Veo 3.1 Fast. Each route has a different starting price and capability fit.
Read the full answer →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.
Read the full answer →Yes on routes that support audio. The current Google video API documents Veo native audio, while current Wan text-to-video models can accept audio and create synchronized video on supported routes.
Read the full answer →The listed Cre8vid starting rates run from $0.025 to $0.069 per generated second, or $0.25 to $0.69 for a ten-second attempt before retries. The exact rate depends on model and settings.
Read the full answer →Primary sources