AI Video & Image Generation API Reference

Call leading image and video models with per-call billing from the same DeepClub balance: OpenAI Images compatible endpoints for images and async multipart tasks for /v1/videos.

Image Generation & Editing API

Use POST https://deepclub.net/v1/images/generations (or client.images.generate) for text-to-image and POST https://deepclub.net/v1/images/edits for reference image editing. Available models include gpt-image-2, gpt-image-2.5-flare, gpt-image-2.5-sunburst, gemini-3-pro-image, gemini-3.1-flash-image, nano-banana-2, and nano-banana-pro.

Async Video Generation API (/v1/videos)

  1. Submit multipart form data to POST https://deepclub.net/v1/videos (model, prompt, seconds) and persist the returned id.
  2. Poll GET https://deepclub.net/v1/videos/{task_id} until status reaches completed or failed.
  3. Download the MP4 via GET https://deepclub.net/v1/videos/{task_id}/content.
curl https://deepclub.net/v1/videos \
  -H "Authorization: Bearer $DEEPCLUB_API_KEY" \
  -F model=Seedance-2.0-Fast \
  -F prompt="Aerial tracking shot over coastal cliffs at golden hour" \
  -F seconds=5

Verify supported size and seconds specifications and live per-call pricing on Live LLM API Pricing.

Frequently Asked Questions

Will polling /v1/videos/{task_id} charge me multiple times?
No. Only creating a new generation task with POST /v1/videos initiates a job. Save the returned task_id and poll GET /v1/videos/{task_id} safely.
What video durations are supported by Seedance 2.0 and MiniMax-H3?
Seedance 2.0 examples use 5 seconds and MiniMax-H3 examples use 10 seconds; always verify exact published specifications per model on /pricing.

Updated