Aliyun Wan I2v

Wan 2.7 Image-to-Video

Validation

mkdir -p output/aliyun-wan-i2v
python -m py_compile skills/ai/video/aliyun-wan-i2v/scripts/generate_i2v.py && echo "py_compile_ok" > output/aliyun-wan-i2v/validate.txt

Pass criteria: command exits 0 and output/aliyun-wan-i2v/validate.txt is generated.

Output And Evidence

• Save task IDs, polling responses, and final video URLs to output/aliyun-wan-i2v/.
• Keep at least one end-to-end run log for troubleshooting.

Prerequisites

• Install SDK (recommended in a venv):

python3 -m venv .venv
. .venv/bin/activate
python -m pip install dashscope

• Set DASHSCOPE_API_KEY in your environment, or add dashscope_api_key to ~/.alibabacloud/credentials.

Critical model names

• wan2.7-i2v — supports first-frame, first+last frame, video continuation, and audio-driven generation

Capabilities

API endpoint (async only)

POST https://dashscope.aliyuncs.com/api/v1/services/aigc/video-generation/video-synthesis

Required headers:

• Authorization: Bearer $DASHSCOPE_API_KEY
• Content-Type: application/json
• X-DashScope-Async: enable

Singapore endpoint: replace dashscope.aliyuncs.com with dashscope-intl.aliyuncs.com.

Normalized interface

Request

• prompt (string, optional) — up to 5000 characters, describes desired video content
• negative_prompt (string, optional) — up to 500 characters
• media (array, required) — media objects with type and url fields:
• type: first_frame | last_frame | driving_audio | first_clip
• url: public URL (HTTP/HTTPS) or OSS temporary URL
• resolution (string, optional) — 720P or 1080P (default: 1080P)
• duration (integer, optional) — video length in seconds, range [2, 15] (default: 5)
• prompt_extend (boolean, optional) — AI prompt rewriting (default: true)
• watermark (boolean, optional) — add "AI generated" watermark (default: false)
• seed (integer, optional) — range [0, 2147483647]

Media input limits

Images (first_frame, last_frame):

• Formats: JPEG, JPG, PNG (no transparency), BMP, WEBP
• Resolution: [240, 8000] pixels per side
• Aspect ratio: 1:8 to 8:1
• Max size: 20MB

Audio (driving_audio):

• Formats: wav, mp3
• Duration: 2-30s
• Max size: 15MB
• Auto-truncated to duration value if longer

Video (first_clip):

• Formats: mp4, mov
• Duration: 2-10s
• Resolution: [240, 4096] pixels per side
• Aspect ratio: 1:8 to 8:1
• Max size: 100MB

Response (task creation)

• output.task_id (string) — use for polling, valid 24 hours
• output.task_status (string) — PENDING | RUNNING | SUCCEEDED | FAILED | CANCELED
• request_id (string)

Response (task result)

• output.video_url (string) — generated video URL
• output.orig_prompt (string) — original prompt
• output.actual_prompt (string) — rewritten prompt (if prompt_extend enabled)
• usage.video_count (integer)
• usage.video_duration (integer) — duration in seconds

Quick start (Python + HTTP)

import os
import json
import time
import requests

API_KEY = os.getenv("DASHSCOPE_API_KEY")
BASE_URL = "https://dashscope.aliyuncs.com/api/v1"

def create_i2v_task(req: dict) -> str:
    """Create an image-to-video task and return task_id."""
    payload = {
        "model": "wan2.7-i2v",
        "input": {
            "prompt": req.get("prompt", ""),
            "media": req["media"],
        },
        "parameters": {
            "resolution": req.get("resolution", "1080P"),
            "duration": req.get("duration", 5),
            "prompt_extend": req.get("prompt_extend", True),
            "watermark": req.get("watermark", False),
        },
    }
    if req.get("negative_prompt"):
        payload["input"]["negative_prompt"] = req["negative_prompt"]
    if req.get("seed") is not None:
        payload["parameters"]["seed"] = req["seed"]

    resp = requests.post(
        f"{BASE_URL}/services/aigc/video-generation/video-synthesis",
        headers={
            "Authorization": f"Bearer {API_KEY}",
            "Content-Type": "application/json",
            "X-DashScope-Async": "enable",
        },
        json=payload,
    )
    resp.raise_for_status()
    data = resp.json()
    return data["output"]["task_id"]


def poll_task(task_id: str, interval: int = 15) -> dict:
    """Poll until task completes. Returns final response."""
    while True:
        resp = requests.get(
            f"{BASE_URL}/tasks/{task_id}",
            headers={"Authorization": f"Bearer {API_KEY}"},
        )
        resp.raise_for_status()
        data = resp.json()
        status = data["output"]["task_status"]
        if status in ("SUCCEEDED", "FAILED", "CANCELED"):
            return data
        time.sleep(interval)

Media combination examples

# First-frame only
media = [{"type": "first_frame", "url": "https://example.com/image.jpg"}]

# First + last frame interpolation
media = [
    {"type": "first_frame", "url": "https://example.com/start.jpg"},
    {"type": "last_frame", "url": "https://example.com/end.jpg"},
]

# Audio-driven from first frame
media = [
    {"type": "first_frame", "url": "https://example.com/face.jpg"},
    {"type": "driving_audio", "url": "https://example.com/speech.mp3"},
]

# Video continuation
media = [{"type": "first_clip", "url": "https://example.com/clip.mp4"}]

Error handling

Output location

• Default output: output/aliyun-wan-i2v/videos/
• Override base dir with OUTPUT_DIR.

Anti-patterns

• Do not use model names other than wan2.7-i2v.
• Do not call this API synchronously — async header is required.
• Do not pass duplicate media types (e.g., two first_frame entries).
• Video URLs expire after 24 hours; download and persist immediately.
• Do not use this API for video editing — use aliyun-wan-videoedit instead.

Workflow

1) Confirm user intent: first-frame, first+last frame, video continuation, or audio-driven. 2) Prepare media array with correct types and valid URLs. 3) Create async task and poll for results. 4) Download and save generated video before URL expiration.

References

• See references/api_reference.md for full HTTP API details.
• See references/sources.md for source links.

DashScope API Reference (Wan 2.7 Image-to-Video)

Model

Endpoint

POST https://dashscope.aliyuncs.com/api/v1/services/aigc/video-generation/video-synthesis

Singapore: https://dashscope-intl.aliyuncs.com/api/v1/services/aigc/video-generation/video-synthesis

Required headers

Request body

{
  "model": "wan2.7-i2v",
  "input": {
    "prompt": "描述文本",
    "negative_prompt": "不希望出现的内容",
    "media": [
      {"type": "first_frame", "url": "https://..."},
      {"type": "driving_audio", "url": "https://..."}
    ]
  },
  "parameters": {
    "resolution": "1080P",
    "duration": 5,
    "prompt_extend": true,
    "watermark": false,
    "seed": 42
  }
}

input fields

media types

Media combinations

parameters fields

Task creation response

{
  "output": {
    "task_status": "PENDING",
    "task_id": "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
  },
  "request_id": "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
}

Task polling

GET https://dashscope.aliyuncs.com/api/v1/tasks/{task_id}
Header: Authorization: Bearer $DASHSCOPE_API_KEY

Success response

{
  "request_id": "...",
  "output": {
    "task_id": "...",
    "task_status": "SUCCEEDED",
    "video_url": "https://...",
    "orig_prompt": "original prompt text",
    "actual_prompt": "rewritten prompt text"
  },
  "usage": {
    "video_count": 1,
    "video_duration": 5
  }
}

Task statuses

Sources

• 万相-图生视频2.7 API参考 — Official API documentation
• 视频生成模型概览 — Model overview and selection guide
• 模型价格 — Pricing details
• 上传文件获取临时URL — File upload for OSS temporary URLs
• 文生视频/图生视频 Prompt 指南 — Prompt writing guide