wan-video — Claude Code skill: what it does, how to install
wan-video generates AI videos from text prompts and animates images using Alibaba Wan models through the AceDataCloud API.
Apache-2.0 238 installs ★ 17 updated 2026-10-25
npx skills add AceDataCloud/Skills --skill wan-video -a claude-codeWhat wan-video does
This skill connects your coding assistant to the AceDataCloud video generation endpoints. It supports text-to-video (wan2.6-t2v), image animation (wan2.6-i2v, wan2.6-i2v-flash), and reference video transfer (wan2.6-r2v). The wan3.0-video model accepts mixed media inputs, including reference frames and files.
You can set resolutions between 480P and 1080P, pick clip lengths, or enable background audio. Multi-shot editing is available with shot_type: "multi" when you need multiple angles instead of a single continuous take.
Calls fail immediately without ACEDATACLOUD_API_TOKEN configured in your .env file. Generation tasks run asynchronously, requiring status checks through the /wan/tasks endpoint.
What it helps with
- generate video clips directly from descriptive text prompts
- animate still images into short video scenes
- transfer character features from reference footage to a new scene
- produce multi-shot video cuts with sound
Example prompts for the agent
- Generate a 5-second 720P video of ocean waves crashing on rocks using wan2.6-t2v with audio enabled
- Animate the image at https://example.com/scene.jpg with a slow camera pan using wan2.6-i2v-flash
- Create a 10-second multi-cut clip of a chef cooking in a busy kitchen with wan2.6-t2v
How to install wan-video
1Skills CLI (Vercel)
npx skills add AceDataCloud/Skills --skill wan-video -a claude-code2Manually: put the files in place
~/.claude/skills/wan-video/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/wan-video/
1Skills CLI (Vercel)
npx skills add AceDataCloud/Skills --skill wan-video -a codex2Manually: put the files in place
.agents/skills/wan-video/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add AceDataCloud/Skills --skill wan-video -a cursor2Manually: put the files in place
.cursor/skills/wan-video/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install AceDataCloud/Skills skills/wan-video2Manually: put the files in place
.github/skills/wan-video/SKILL.md1Skills CLI (Vercel)
npx skills add AceDataCloud/Skills --skill wan-video -a gemini-cli2Platform CLI
gemini skills install https://github.com/AceDataCloud/Skills --path skills/wan-videoManually: put the files in place
.devin/skills/wan-video/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add AceDataCloud/Skills --skill wan-video -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about wan-video
How to install wan-video in Claude Code?
Run this command in your terminal: npx skills add AceDataCloud/Skills --skill wan-video -a claude-code.
What does wan-video do?
It lets your AI assistant call AceDataCloud endpoints to generate videos via Alibaba Wan models from prompts, pictures, or reference clips.
What credentials are required before running wan-video?
You must set ACEDATACLOUD_API_TOKEN in your project .env file to authenticate calls to the API.
Source text
SKILL.md · Apache-2.0
---
name: wan-video
description: Generate AI videos with Wan (Alibaba) via AceDataCloud API. Use when creating videos from text prompts or animating images into video. Supports text-to-video, image-to-video, reference video transfer, multi-resolution (480P-1080P), and optional audio.
license: Apache-2.0
metadata:
author: acedatacloud
version: "1.0"
compatibility: Requires ACEDATACLOUD_API_TOKEN in .env file (see _shared/authentication.md). Optionally pair with mcp-wan for tool-use.
---
# Wan Video Generation
Generate AI videos through AceDataCloud's Wan (Alibaba) API.
> **Setup:** See [authentication](../_shared/authentication.md) for token setup.
## Quick Start
```bash
curl -X POST https://api.acedata.cloud/wan/videos \
-H "Authorization: Bearer $ACEDATACLOUD_API_TOKEN" \
-H "Content-Type: application/json" \
-d '{"action": "text2video", "prompt": "a dolphin jumping through ocean waves at golden hour", "model": "wan2.6-t2v"}'
```
> **Async:** See [async task polling](../_shared/async-tasks.md). Poll via `POST /wan/tasks` with `{"id": "..."}`.
## Models
| Model | Type | Best For |
|-------|------|----------|
| `wan2.6-t2v` | Text-to-Video | Creating video from text description |
| `wan2.6-i2v` | Image-to-Video | Animating a still image into video |
| `wan2.6-r2v` | Reference Video-to-Video | Character extraction and transfer from reference video |
| `wan2.6-i2v-flash` | Image-to-Video (Fast) | Quick image-to-video generation |
| `wan3.0-video` | All-in-One | Text, frames, reference media, files, and public links |
## Workflows
### 1. Text-to-Video
```json
POST /wan/videos
{
"action": "text2video",
"prompt": "a time-lapse of flowers blooming in a meadow",
"model": "wan2.6-t2v",
"resolution": "720P",
"duration": 5
}
```
### 2. Image-to-Video
Animate a still image into a video clip.
```json
POST /wan/videos
{
"action": "image2video",
"prompt": "gentle wind blows through the scene",
"model": "wan2.6-i2v",
"image_url": "https://example.com/landscape.jpg",
"resolution": "720P",
"duration": 5
}
```
### 3. Image-to-Video (Flash)
Faster image-to-video generation with reduced latency.
```json
POST /wan/videos
{
"action": "image2video",
"prompt": "camera slowly pans across the landscape",
"model": "wan2.6-i2v-flash",
"image_url": "https://example.com/scene.jpg"
}
```
### 4. Reference Video Transfer
Extract characters or timbres from a reference video and transfer them into a new generation.
```json
POST /wan/videos
{
"action": "text2video",
"prompt": "the character walks through a futuristic city at night",
"model": "wan2.6-r2v",
"reference_video_urls": ["https://example.com/reference.mp4"]
}
```
### 5. Multi-Cut Editing
Generate a video with multiple shots rather than a single continuous take.
```json
POST /wan/videos
{
"action": "text2video",
"prompt": "a chef preparing a meal in a busy kitchen",
"model": "wan2.6-t2v",
"shot_type": "multi",
"duration": 10
}
```
### 6. Video with Audio
Enable audio generation alongside the video.
```json
POST /wan/videos
{
"action": "text2video",
"prompt": "ocean waves crashing on a rocky shore",
"model": "wan2.6-t2v",
"audio": true
}
```
## Parameters
| Parameter | Required | Values | Description |
|-----------|----------|--------|-------------|
| `action` | Yes | `"text2video"`, `"image2video"` | Action type |
| `prompt` | Yes | string | Scene description |
| `model` | Yes | `"wan2.6-t2v"`, `"wan2.6-i2v"`, `"wan2.6-r2v"`, `"wan2.6-i2v-flash"` | Model |
| `image_url` | For image2video | string | Source image URL (required for image-to-video) |
| `negative_prompt` | No | string (max 500 chars) | Content to exclude from generation |
| `reference_video_urls` | For r2v | array of strings | Reference videos for character/timbre extraction |
| `shot_type` | No | `"single"`, `"multi"` | Continuous shot or multi-cut editing |
| `audio` | No | boolean | Enable audio in the generated video |
| `audio_url` | No | string | Reference audio URL |
| `resolution` | No | `"480P"`, `"720P"`, `"1080P"` | Output resolution (default: 720P) |
| `size` | No | string | The size of the generated video |
| `duration` | No | `5`, `10`, `15` | Video duration in seconds |
| `prompt_extend` | No | boolean | Enable LLM-based prompt rewriting |
| `callback_url` | No | string | Async webhook notification URL |
## Gotchas
- `image_url` is **required** for `wan2.6-i2v` and `wan2.6-i2v-flash` models
- `reference_video_urls` is used only with `wan2.6-r2v` for character/timbre transfer
- `negative_prompt` has a maximum length of 500 characters
- Supported durations are 5, 10, or 15 seconds only
- Default resolution is 720P; use 1080P for higher quality at increased cost
- `shot_type: "multi"` produces multi-cut edits rather than a single continuous shot
> **MCP:** `pip install mcp-wan` | Hosted: `https://wan.mcp.acedata.cloud/mcp` | See [all MCP servers](../_shared/mcp-servers.md)
## Wan 3 All-in-One
`wan3.0-video` infers the workflow from `media`. Supported types are `first_frame`, `last_frame`, `reference_image` (up to 10), `reference_video` (up to 5, 15 seconds total), `reference_audio` (up to 5, 15 seconds total), `file`, and `link`. Frame mode cannot be mixed with reference/file/link mode.
```json
POST /wan/videos
{
"model": "wan3.0-video",
"prompt": "Use image 1 to create a cinematic product video",
"media": [{"type":"reference_image","url":"https://cdn.acedata.cloud/r9vsv9.png"}],
"resolution": "720P",
"ratio": "16:9",
"duration": 5,
"audio": true,
"async": true
}
```
Wan 3 supports 2–30 seconds or `-1` automatic duration. Poll with the existing `/wan/tasks` endpoint.
© acedatacloud · Apache-2.0 · AceDataCloud/Skills