Wan 3.0

Generate videos from text, frames, or multimodal references.

POST

/api/v1/generations

Example request

Set your FLUX_API_KEY before running. See the parameters below for supported settings.

cURL

curl https://api.flux-context.org/api/v1/generations \
  -H "Authorization: Bearer $FLUX_API_KEY" \
  -H "Content-Type: application/json" \
  --data '{
  "model": "wan3",
  "input": {
    "mode": "text-to-video",
    "prompt": "A calm seaside sunrise with gentle camera movement",
    "resolution": "480p",
    "duration": 4,
    "sound": false,
    "aspect_ratio": "16:9"
  }
}'

Parameters

Set model to wan3. These fields belong inside input.

Parameter	Description
`resolution` string · Optional	Output resolution tier. Values are case-sensitive. `480p720p1080p` Default: `720p`
`aspect_ratio` string · Optional	Output proportions; see mode constraints below. `auto16:94:31:13:49:16` Default: `auto`
`sound` boolean · Optional	Generate sound with the video. This is not a reference audio input. Default: `true`
`watermark` boolean · Optional	Whether to request a watermark. Default: `false`
`mode` string · Required	Choose a supported generation mode. `text-to-videoimage-to-videoreference-to-video`
`prompt` string · Optional	Describe the desired output. Maximum 5,000 characters. Default:
`image_urls` string[] · Optional	Public HTTPS reference images. Video image order and count depend on mode. Up to 10 items.
`video_urls` string[] · Optional	Public HTTPS MP4 or MOV references. Flux measures duration before charging. Maximum file size 200 MiB. Up to 5 items.
`audio_urls` string[] · Optional	Public HTTPS MP3, WAV or M4A references. Flux measures duration before charging. Maximum file size 15 MiB. Audio is not billed separately. Up to 5 items.
`reference_intent` string · Optional	Generate new content or edit a reference video. See model constraints for duration behavior. `generateedit` Default: `generate`
`duration` integer · Optional	Output duration in seconds. See model constraints for editing and automatic duration. Range: 2–30. Default: `5`
`seed` integer · Optional	Optional random seed; repeatability is not guaranteed. Range: 0–2147483647.

Reference media

Use your own publicly accessible HTTPS media URLs. Links must remain valid throughout queueing and generation. Base64 and file upload bodies are not supported by this endpoint.

image-to-video

cURL

curl https://api.flux-context.org/api/v1/generations \
  -H "Authorization: Bearer $FLUX_API_KEY" \
  -H "Content-Type: application/json" \
  --data '{
  "model": "wan3",
  "input": {
    "mode": "image-to-video",
    "prompt": "A calm seaside sunrise with gentle camera movement",
    "resolution": "480p",
    "duration": 4,
    "sound": false,
    "aspect_ratio": "auto",
    "image_urls": [
      "https://example.com/front.jpg",
      "https://example.com/back.jpg"
    ]
  }
}'

reference-to-video

cURL

curl https://api.flux-context.org/api/v1/generations \
  -H "Authorization: Bearer $FLUX_API_KEY" \
  -H "Content-Type: application/json" \
  --data '{
  "model": "wan3",
  "input": {
    "mode": "reference-to-video",
    "prompt": "A calm seaside sunrise with gentle camera movement",
    "resolution": "480p",
    "duration": 4,
    "sound": false,
    "aspect_ratio": "auto",
    "image_urls": [
      "https://example.com/front.jpg",
      "https://example.com/back.jpg"
    ]
  }
}'

Video and audio references

cURL

curl https://api.flux-context.org/api/v1/generations \
  -H "Authorization: Bearer $FLUX_API_KEY" \
  -H "Content-Type: application/json" \
  --data '{
  "model": "wan3",
  "input": {
    "mode": "reference-to-video",
    "prompt": "A calm seaside sunrise with gentle camera movement",
    "resolution": "480p",
    "duration": 4,
    "sound": false,
    "aspect_ratio": "auto",
    "image_urls": [
      "https://example.com/front.jpg"
    ],
    "video_urls": [
      "https://example.com/reference.mp4"
    ],
    "audio_urls": [
      "https://example.com/reference.mp3"
    ]
  }
}'

Edit a video

cURL

curl https://api.flux-context.org/api/v1/generations \
  -H "Authorization: Bearer $FLUX_API_KEY" \
  -H "Content-Type: application/json" \
  --data '{
  "model": "wan3",
  "input": {
    "mode": "reference-to-video",
    "prompt": "Edit the video to use a warm sunset background",
    "resolution": "480p",
    "duration": 4,
    "sound": false,
    "aspect_ratio": "auto",
    "image_urls": [
      "https://example.com/front.jpg"
    ],
    "video_urls": [
      "https://example.com/reference.mp4"
    ],
    "audio_urls": [
      "https://example.com/reference.mp3"
    ],
    "reference_intent": "edit"
  }
}'

Credits

Resolution	Credits / second
480p	10
720p	18
1080p	36

Video reference seconds are included in the billable duration. See the model-specific rounding and editing rules below. Sound does not change the price. Confirmed failed requests are refunded.

Credits = ceil((fixed output seconds + measured input video seconds, rounded to 3 decimals) × rate). Reference audio is not billed separately. Editing derives a fixed output duration before charging.

Constraints

A prompt is required for text-to-video and optional with media. Image-to-video accepts 1–2 frames. Reference mode supports up to 10 images, 5 videos and 5 audio files.

Each video/audio must be 1–15 seconds; video total and audio total must each be at most 15 seconds. Output plus input video duration must not exceed 30 seconds.

With reference_intent edit, a prompt is required and aspect_ratio is set to auto. Flux derives a fixed output duration: min(max(2, ceil(input video seconds + 1)), floor(30 - input video seconds)); the supplied duration is replaced. Automatic duration -1, file_urls and link_urls are not supported.

Use returned dimensions when available. Unverified video dimensions are null. Videos are delivered as MP4. Reference and editing rules depend on the model.

Response

A successful submission returns HTTP 202 and a request ID. Response excerpt:

JSON

{
  "id": "REQUEST_ID",
  "model": "wan3",
  "status": "processing"
}

Continue readingPoll & download