RelayDanceOne API.
Cinema-grade AI video.
Generate video with Seedance 2.0 and HappyHorse through one OpenAI-compatible endpoint, plus gpt-image-2 image generation. Submit a prompt, get a finished clip: up to 15 seconds, audio included, reference images and faces supported. No queue, no KYC.
Video and image models from leading AI labs
# 1) Submit a video generation task
curl https://relaydance.com/v1/video/generations \
-H "Authorization: Bearer $RELAYDANCE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "doubao-seedance-2-0-1080p",
"prompt": "A dancer spins in golden hour light, cinematic",
"seconds": "10",
"metadata": {
"ratio": "16:9",
"resolution": "1080p",
"generate_audio": true,
"callback_url": "https://your-server.example.com/webhook"
}
}'
# -> { "task_id": "...", "status": "submitted" }
# 2) Poll (or skip polling by using metadata.callback_url)
curl https://relaydance.com/v1/video/generations/TASK_ID \
-H "Authorization: Bearer $RELAYDANCE_API_KEY"Video API
Production video, two HTTP calls.
Submit a generation task, receive a finished clip. OpenAI-compatible request shapes, asynchronous task workflow, webhook callback support.
- ✓Seedance 2.0 and HappyHorse: text-to-video and image-to-video up to 15s with audio, 480p to 4K
- ✓gpt-image-2 image generation with free image output, plus multi-reference fusion (9 images, 3 videos, 3 audio)
- ✓No queue, no KYC, face and portrait generation supported
- ✓Special SKUs beyond the official platform, pay with USDT or card, failed requests never billed
Integrate in minutes
Two HTTP calls to a finished clip.
Submit a task, then poll it, or set metadata.callback_url and receive a webhook instead.
# 1) Submit a video generation task
curl https://relaydance.com/v1/video/generations \
-H "Authorization: Bearer $RELAYDANCE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "doubao-seedance-2-0-1080p",
"prompt": "A dancer spins in golden hour light, cinematic",
"seconds": "10",
"metadata": {
"ratio": "16:9",
"resolution": "1080p",
"generate_audio": true,
"callback_url": "https://your-server.example.com/webhook"
}
}'
# -> { "task_id": "...", "status": "submitted" }
# 2) Poll (or skip polling by using metadata.callback_url)
curl https://relaydance.com/v1/video/generations/TASK_ID \
-H "Authorization: Bearer $RELAYDANCE_API_KEY"Webhook callbacks
Skip polling. Get a webhook.
Set metadata.callback_url when you submit. When the clip finishes, we POST the final task state to your server, with retries.
POST <your callback_url>
{
"task_id": "cgt-2026...",
"status": "succeeded",
"url": "https://.../output.mp4",
"format": "mp4",
"metadata": { "duration": 10.0, "resolution": "1080p" }
}Compatibility
Works with your existing stack.
Plain HTTP plus webhooks. Use any language or automation tool.
Use cases
Built for video products.
Short-form content
Generate vertical clips for TikTok, Reels and Shorts pipelines. Batch-render variations from one prompt template.
E-commerce video
Turn product stills into motion: image-to-video with reference images keeps the product faithful while adding cinematic movement.
Apps & agents
Add video generation to your product with two HTTP calls: submit and poll, or just set a callback URL and wait for the webhook.
Why RelayDance
Everything you need, nothing you don't.
Seedance 2.0 & HappyHorse video
Text-to-video and image-to-video at 480p to native 4K, up to 15 seconds, optional generated audio. No queue, tasks start immediately.
gpt-image-2 images
OpenAI gpt-image-2 image generation through /v1/images/generations. Image output is free, you pay only for input.
Async by design
Submit returns a task id immediately. Poll the task, or set metadata.callback_url and receive a webhook when the clip is ready.
Multi-reference fusion
Blend up to 9 reference images, 3 reference videos and 3 audio tracks in one generation request.
USDT and card payments
Top up with USDT or pay by credit and debit card via Stripe checkout.
Zero failed-request charges
Failed generations are never billed. You only pay for successfully delivered output.
FAQ