| Provider | Status | Price/sec | API | Webhooks | Batch | Failover | Models |
|---|---|---|---|---|---|---|---|
RecommendedCreativeAI | Live | $0.02–0.09 | Kling v3Seedance 1.5ViduWan 2.1Veo 3.1GPT Image 1 | ||||
Kling (Kuaishou) | Live | $0.153 | Kling v3Kling O3 | ||||
Seedance (ByteDance) | Live | $0.06 | Seedance 1.5 | ||||
Vidu | Live | $0.05 | Vidu | ||||
Wan (Alibaba) | Live | $0.018 | Wan 2.1 | ||||
Veo (Google) | Live | $0.09 | Veo 3.1 | ||||
Sora (OpenAI) | Shut Down | N/A | Sora 2 (subscription) |
Pricing accurate as of April 2026. Contact providers directly for enterprise rates.
Kling, Seedance, Vidu, Wan, Veo, and more. Single API key, same endpoint.
HMAC-signed callbacks with 3 retry attempts (0s, 5s, 30s). No polling loops.
Submit hundreds of video jobs in one request. Track batch status via API.
If one provider fails, requests auto-route to backup. No duplicate charge.
From ~$0.02/second for Wan to ~$0.09/second for Veo. No subscription required.
Change base_url and api_key. Your existing Python/Node code works as-is.
# Python — Same code for all video models
from openai import OpenAI
client = OpenAI(
api_key="your-creativeai-key",
base_url="https://api.creativeai.run/v1",
)
# Text-to-video
response = client.post("/video/generations", body={
"model": "kling-v3", # or seedance-1.5, vidu, wan-2.1, veo-3.1
"prompt": "Drone shot over coastal city at golden hour",
"duration": "5",
"aspect_ratio": "16:9",
"webhook_url": "https://yourapp.com/webhooks/video",
}, cast_to=object)
# Webhook payload includes: job_id, status, output_url, failover_usedDirect APIs require polling for completion, have no failover, and need separate integrations for each model. CreativeAI gives you all models through one API key, with webhook delivery, batch processing, and automatic multi-model failover. If Kling is down, your request routes to Seedance automatically.
For synchronous failures (4xx/5xx at submission), you get an immediate error and can retry. For async failures (provider crashes mid-render), we automatically retry with a backup model if failover is enabled. No duplicate charge for failed generations.
When your video is ready, we POST a signed payload to your webhook_url with the job_id, status, and output_url. HMAC-SHA256 signature lets you verify authenticity. Three delivery attempts with exponential backoff (0s, 5s, 30s). If all fail, the result stays available via the status API.
Yes. CreativeAI is a standard REST API with Bearer token auth. Use HTTP Request nodes (n8n) or HTTP modules (Make/Zapier). We have tutorials for all three platforms: `/tutorials/n8n-integration` and `/tutorials/make-zapier-integration`.
Sora 2 is only available through ChatGPT Pro ($200/month subscription). There is no standalone API. OpenAI shut down the original Sora API on March 25, 2026. For programmatic video generation, you need an alternative.