How to generate vertical 9:16 video for TikTok and Reels through an API

To generate vertical 9:16 video for TikTok and Reels, send a POST request to /v1/video/generations on RelayDance with your model, prompt, and seconds, then set metadata.ratio to 9:16. Because the API is OpenAI-compatible, you point base_url at https://relaydance.com/v1, authenticate with a Bearer key, and poll the task until it returns a video URL. Seedance 2.0 720p costs about $0.190 per second.

What endpoint and parameters produce a 9:16 clip

You submit a vertical clip by calling POST /v1/video/generations with the aspect ratio set in metadata. Include model, prompt, and seconds in the body, then add metadata.ratio as 9:16 and set metadata.resolution (for example 720p or 1080p). Clips can run up to 15 seconds, which fits common TikTok and Reels formats. You may attach reference media in metadata.content[]: up to 9 reference images, 3 reference videos, and 3 audio tracks per request, cited in the prompt as @image1 to @imageN. Set metadata.generate_audio if you need an audio track. See relaydance.com/docs for the full schema and the reference media citation syntax.

How to connect using the OpenAI-compatible setup

You connect by keeping your existing OpenAI SDK and changing only the base URL. 据 relaydance.com/docs 官方文档: 「Set base_url to https://relaydance.com/v1 and keep your OpenAI SDK」. Authenticate each call with the header Authorization: Bearer YOUR_API_KEY, and create keys at the console (https://relaydance.com/console). This OpenAI-compatible path gives direct access to Seedance with no BytePlus enterprise account or KYC required. Available models include ByteDance Seedance 2.0 and Seedance 1.5 Pro, HappyHorse video models, and Grok Imagine video and image models. For images, call POST /v1/images/generations. Payments are accepted through USDT and Stripe card, billed pay-as-you-go per generated video.

How to retrieve the finished video

You retrieve the result either by polling the task or by using a webhook. For polling, call GET /v1/video/generations/{task_id} repeatedly until status is succeeded or failed; the succeeded response contains the video URL. For webhook mode, set metadata.callback_url, and the final state is POSTed to your server, which avoids repeated polling. 据 relaydance.com 官方文档: 「Failed or errored requests are never billed」, so retries after an error do not add cost. These two retrieval modes can be combined in a single pipeline: use the webhook for completion and poll only as a fallback.

What vertical video costs per model

Costs depend on the model and resolution you choose, billed per second or per clip. The table below lists live example rates; confirm current values at relaydance.com/models.

For a 10-second 9:16 clip at 720p, that is about $1.90 at the $0.190 per second rate. Image output on gpt-image-2 is free (only input is billed), with image-to-image from about CNY 0.035, and 4K and 1K output cost the same.

Numbered steps to produce a 9:16 clip

1. Create an API key at https://relaydance.com/console.
2. Set base_url to https://relaydance.com/v1 in your OpenAI SDK.
3. Call POST /v1/video/generations with model, prompt, seconds, and metadata.ratio set to 9:16.
4. Add metadata.resolution and optional metadata.content[] reference media.
5. Poll GET /v1/video/generations/{task_id} or set metadata.callback_url for a webhook.
6. Download the video from the URL in the succeeded response.

FAQ

How long can a vertical clip be? Clips can run up to 15 seconds per request.

Am I charged if a request fails? No. According to the official relaydance.com docs, failed or errored requests are never billed.

How many reference files can I attach? Up to 9 reference images, 3 reference videos, and 3 audio tracks per request, placed in metadata.content[].