Video generation

Use /v1/videos/generations for video creation.

curl https://apiany.ai/v1/videos/generations \
  -H "Authorization: Bearer $APIANY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2.0",
    "prompt": "A cinematic product launch shot for an AI API platform",
    "duration": "10s",
    "resolution": "720p",
    "ratio": "16:9"
  }'

APIAny schedules processing immediately after the task record is committed. The 202 response does not wait for video generation to finish, so poll the returned task until it reaches a terminal state. The periodic task worker remains a recovery mechanism for pending tasks.

Prompt safety

Prompts that clearly combine a minor with exposure or emphasis of sensitive body areas, or with explicit sexual or nudity content, return HTTP 400 with code content_policy. Validation runs before task creation and credit reservation, so no credits are charged. Revise the prompt to use an adult subject (18+), remove sensitive-body close-ups, and keep the subject fully clothed. Ordinary child-safe family, classroom, sports, and art scenes are not blocked by this targeted rule.

Wan 3.0 Prime requests

Wan 3.0 Prime uses three public model IDs with separate input contracts:

Model ID	Input
`wan3.0-prime-text-to-video`	Prompt only. Reference media is rejected.
`wan3.0-prime-image-to-video`	Required `image_start`; optional `image_end`.
`wan3.0-prime-reference-video`	Ordered `image_urls`, `video_urls`, and `audio_urls`, plus optional document or web context.

All three accept an integer duration from 2 to 30, quality of 480p, 720p, or 1080p, and aspect_ratio of adaptive, 16:9, 9:16, 1:1, 4:3, or 3:4. generate_audio defaults to true. resolution remains an alias for quality. Automatic duration, task cancellation, negative_prompt, and prompt_extend are not supported on these public model IDs.

Text to video

{
  "model": "wan3.0-prime-text-to-video",
  "prompt": "A precise product film with a slow camera orbit",
  "duration": 5,
  "quality": "720p",
  "aspect_ratio": "16:9",
  "generate_audio": true
}

First and last frame

{
  "model": "wan3.0-prime-image-to-video",
  "prompt": "Move smoothly from the opening composition to the final frame",
  "image_start": "https://assets.example/start.webp",
  "image_end": "https://assets.example/end.webp",
  "duration": 5,
  "quality": "720p",
  "aspect_ratio": "16:9"
}

The image-to-video route does not accept image_urls; use image_start and the optional image_end fields shown above.

Multimodal references

{
  "model": "wan3.0-prime-reference-video",
  "prompt": "Use Image 1 for product design and Video 1 for camera rhythm",
  "image_urls": ["https://assets.example/product.webp"],
  "video_urls": ["https://assets.example/camera-reference.mp4"],
  "audio_urls": ["https://assets.example/ambience.mp3"],
  "duration": 5,
  "quality": "720p",
  "aspect_ratio": "adaptive"
}

The reference route accepts up to 10 images, 5 videos, and 5 audio tracks. Add one document with model_params.file_url or one web page with model_params.link_url; those two fields are mutually exclusive. Numbering is per media type and follows array order: Image 1, Video 1, and Audio 1 refer to the first item of their respective arrays. Save the task ID and use polling or an HTTPS callback. Persist successful outputs promptly because generated result URLs are temporary.

Sync variant

/v1/videos/generations/sync waits for the task result in the same HTTP request. Long-running jobs should use async creation and task polling.

Get started

Integrations

Core APIs

Operations

Prompt safety

Wan 3.0 Prime requests

Text to video

First and last frame

Multimodal references

Sync variant

​Prompt safety

​Wan 3.0 Prime requests

​Text to video

​First and last frame

​Multimodal references

​Sync variant

Prompt safety

Wan 3.0 Prime requests

Text to video

First and last frame

Multimodal references

Sync variant