curl --request POST \
  --url https://api.evolink.ai/v1/videos/generations \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "wan2.7-image-to-video",
  "generation_mode": "first_frame",
  "prompt": "A cat playing piano",
  "image_start": "https://example.com/first_frame.jpg"
}
'

{
  "created": 1757169743,
  "id": "task-unified-1757169743-7cvnl5zw",
  "model": "wan2.7-image-to-video",
  "object": "video.generation.task",
  "progress": 0,
  "status": "pending",
  "task_info": {
    "can_cancel": true,
    "estimated_time": 120
  },
  "type": "video",
  "usage": {
    "billing_rule": "per_call",
    "credits_reserved": 5,
    "user_group": "default"
  }
}

Wan2.7

Wan2.7 Image to Video

WAN2.7 (wan2.7-image-to-video) model supports image-to-video generation with multimodal inputs (image / audio / video)
Choose one of three generation modes via the generation_mode parameter:
- first_frame: First-frame to video — generate a video starting from the given first frame, optional driving audio
- first_last_frame: First-and-last-frame to video — generate a video by interpolating between the first and last frames, optional driving audio
- video_continuation: Video continuation — continue an input video clip; an optional ending frame is allowed (no driving audio)
generation_mode is optional (backward compatible); if omitted, an appropriate mode will be selected automatically based on the materials in the request
Valid material combinations (any other combination will be rejected):
1. image_start (first frame)
2. image_start + audio_urls (first frame + driving audio)
3. image_start + image_end (first + last frame)
4. image_start + image_end + audio_urls (first + last frame + driving audio)
5. video_urls (video continuation)
6. video_urls + image_end (video continuation + last frame)
Asynchronous processing mode, use the returned task ID to query status
Generated video links are valid for 24 hours, please save them promptly

POST

videos

generations

curl --request POST \
  --url https://api.evolink.ai/v1/videos/generations \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "wan2.7-image-to-video",
  "generation_mode": "first_frame",
  "prompt": "A cat playing piano",
  "image_start": "https://example.com/first_frame.jpg"
}
'

{
  "created": 1757169743,
  "id": "task-unified-1757169743-7cvnl5zw",
  "model": "wan2.7-image-to-video",
  "object": "video.generation.task",
  "progress": 0,
  "status": "pending",
  "task_info": {
    "can_cancel": true,
    "estimated_time": 120
  },
  "type": "video",
  "usage": {
    "billing_rule": "per_call",
    "credits_reserved": 5,
    "user_group": "default"
  }
}

Authorizations

Authorization

string

header

required

All APIs require Bearer Token authentication

Get your API Key:

Visit the API Key management page to obtain your API Key

Add to request headers:

Authorization: Bearer YOUR_API_KEY

Body

application/json

model

enum<string>

required

Model name, must be wan2.7-image-to-video

Available options:

wan2.7-image-to-video

Example:

"wan2.7-image-to-video"

generation_mode

enum<string>

Generation mode that determines which material combinations are valid. Explicitly specifying it is recommended

Values:

first_frame: First-frame to video. Required: image_start. Optional: audio_urls. Not accepted: image_end, video_urls
first_last_frame: First-and-last-frame to video. Required: image_start + image_end. Optional: audio_urls. Not accepted: video_urls
video_continuation: Video continuation. Required: video_urls[0]. Optional: image_end (used as ending frame). Not accepted: image_start, audio_urls

Backward-compatible behavior: when generation_mode is omitted, an appropriate mode will be selected automatically based on the materials in the request; explicit specification is recommended to avoid ambiguity

Available options:

first_frame,

first_last_frame,

video_continuation

Example:

"first_frame"

prompt

string

Text prompt for video generation. Supports both Chinese and English; each character/letter counts as 1, with overflow auto-truncated. Maximum length: 5000 characters

Maximum string length: 5000

Example:

"A cat playing piano"

negative_prompt

string

Negative prompt describing what should not appear in the video. Supports both Chinese and English. Maximum length 500 characters; overflow is auto-truncated

Maximum string length: 500

Example:

"Blurry, low quality"

image_start

string<uri>

First-frame image URL

Mode constraints:

first_frame mode: required
first_last_frame mode: required
video_continuation mode: not allowed

Image limits:

Formats: JPEG, JPG, PNG (transparency not supported), BMP, WEBP
Resolution: width and height in [240, 8000] pixels
Aspect ratio: 1:8 ~ 8:1
File size: up to 20MB

Example:

"https://example.com/first_frame.jpg"

image_end

string<uri>

Ending-frame image URL

Mode constraints:

first_last_frame mode: required
video_continuation mode: optional (acts as the ending frame for the continuation)
first_frame mode: not allowed (use first_last_frame if both first and last frames are needed)

Image limits:

Formats: JPEG, JPG, PNG (transparency not supported), BMP, WEBP
Resolution: width and height in [240, 8000] pixels
Aspect ratio: 1:8 ~ 8:1
File size: up to 20MB

Example:

"https://example.com/last_frame.jpg"

video_urls

string<uri>[]

Video continuation URL array. Only 1 element is supported

Mode constraints:

video_continuation mode: required
first_frame / first_last_frame mode: not allowed
Cannot be combined with audio_urls

Video limits:

Formats: mp4, mov
Duration: 2 ~ 10 seconds (length of the input clip itself)
Resolution: width and height in [240, 4096] pixels
Aspect ratio: 1:8 ~ 8:1
File size: up to 100MB

Continuation duration rules:

duration represents the total final output video length (input clip + model-generated continuation)
Generated continuation length = duration − input video length
duration must be ≥ input video length
Billing is based on the total final output length (i.e. duration)

Examples:

Input clip length	duration	Continuation generated	Final output	Billed
3s	15	12s	15s	15s
5s	10	5s	10s	10s
8s	8	0s (input only)	8s	8s

Example:

["https://example.com/clip.mp4"]

audio_urls

string<uri>[]

Driving audio URL array. Currently supports only 1 element. The model will use this audio as the driving source for video generation (e.g. lip sync, motion alignment)

Mode constraints:

first_frame mode: optional
first_last_frame mode: optional
video_continuation mode: not allowed (cannot be combined with video_urls)

Format requirements:

Supported formats: wav, mp3
Duration range: 2 ~ 30 seconds
File size: up to 15MB

Truncation handling:

If audio length exceeds duration, the first N seconds are extracted and the rest discarded
If audio length is shorter than the video duration, the remaining portion is silent. For example: if audio is 3s and video duration is 5s, the first 3s have sound and the last 2s are silent

Required array length: 1 element

Example:

["https://example.com/audio.mp3"]

quality

enum<string>

default:720p

Video quality, defaults to 720p

Options:

720p: Standard definition, standard price, this is the default
1080p: High definition, higher price

Available options:

720p,

1080p

Example:

"720p"

duration

number

default:5

Video duration in seconds (integer). Range 2 ~ 15, default 5

Meaning:

first_frame / first_last_frame modes: total length of the generated video
video_continuation mode: total length of the final output video (= original input clip + model-generated continuation)

Additional constraints in video_continuation mode:

duration must be ≥ input video length (otherwise an error is returned)
Generated continuation length = duration − input video length
When duration equals the input video length, no continuation is generated and the input clip is returned as-is
See the continuation duration rules and examples in the video_urls field for details

Billing: based on the actual generated video duration

Required range: 2 <= x <= 15

Example:

5

seed

integer

Random seed, defaults to random

Notes:

Range: 1 ~ 2147483647
Fixing the seed reduces variation when iterating on prompts and improves reproducibility

Required range: 1 <= x <= 2147483647

Example:

42

prompt_extend

boolean

default:false

Whether to enable intelligent prompt rewriting. When enabled, a large model will optimize the prompt, which significantly improves results for simple or insufficiently descriptive prompts.

Note: Default is false. Omitting the field or sending false will not trigger rewriting; explicitly send true to enable.

Example:

false

callback_url

string<uri>

HTTPS callback URL for task completion

Callback Timing:

Triggered when task is completed, failed, or cancelled
Sent after billing confirmation

Security Restrictions:

Only HTTPS protocol is supported
Callbacks to internal IP addresses are prohibited (127.0.0.1, 10.x.x.x, 172.16-31.x.x, 192.168.x.x, etc.)
URL length must not exceed 2048 characters

Callback Mechanism:

Timeout: 10 seconds
Up to 3 retries after failure (retries at 1/2/4 seconds after failure)
Callback response format is consistent with the task query API response
2xx status codes are considered successful, other status codes trigger retries

Example:

"https://your-domain.com/webhooks/video-task-completed"

Response

Video task created successfully

created

integer

Task creation timestamp

Example:

1757169743

string

Task ID

Example:

"task-unified-1757169743-7cvnl5zw"

model

string

Actual model name used

Example:

"wan2.7-image-to-video"

object

enum<string>

Specific task type

Available options:

video.generation.task

progress

integer

Task progress percentage (0-100)

Required range: 0 <= x <= 100

Example:

0

status

enum<string>

Task status

Available options:

pending,

processing,

completed,

failed

Example:

"pending"

task_info

object

Detailed video task information

Show child attributes

type

enum<string>

Task output type

Available options:

text,

image,

audio,

video

Example:

"video"

usage

object

Usage and billing information

Show child attributes

Wan2.7 Text to Video Wan2.7 Reference Video