Mcp Happyhorse

io.github.AceDataCloudv2026.10.4.1更新于 Oct 4, 2026

Generate and edit Happy Horse AI videos through Ace Data Cloud

已验证Streamable HTTP可网页运行AI & ML Media & Design

概览

AI 生成的概览

让助手通过 Ace Data Cloud 用文本、图片或参考图生成和编辑 Happy Horse AI 视频。

功能: 封装 Happy Horse 视频 API，提供文生视频、首帧图片动画、使用 1-9 张参考图生成视频，以及最多 5 张参考图的视频编辑工具。还提供单个和批量任务查询工具以及模型列表工具。生成是异步的：助手需保留返回的 task_id 并轮询，直到出现最终 video_url 或错误。输出为 720P 或 1080P，时长 3-15 秒。
适用场景: 适合在对话中让助手生成短视频或编辑已有片段，例如把提示词或静态图片变成视频，或对源视频做风格化处理。也适合已在使用 Ace Data Cloud、希望免去自行对接 API 的用户。
运行要求: 可使用 happyhorse.mcp.acedata.cloud 远程端点，或从 PyPI 安装本地进程（mcp-happyhorse，用 uvx 或 pip 运行）。本地 stdio 模式需要环境变量 ACEDATACLOUD_API_TOKEN；托管端点接受 Bearer 令牌请求头或 OAuth。令牌需从 Ace Data Cloud 平台获取。需要访问 Ace Data Cloud API 的网络。

安装前请注意

需要 ACEDATACLOUD_API_TOKEN 密钥，它用于授权 API 调用，可能在 Ace Data Cloud 账户上产生费用。生成和编辑会把任务提交到第三方服务，提示词、图片和视频都会发送给 Ace Data Cloud。视频生成是异步的并消耗额度或配额，大量使用前请确认计费方式。

安装

在 SourceWeft 中

打开控制台中的 Mcp Happyhorse，将其添加到工作区。
为需要使用其工具的对话启用该服务。

Web executable，通过 Streamable HTTP。远程服务在工作区中配置后即可从网页运行时运行。

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。

{
  "mcpServers": {
    "mcp-happyhorse": {
      "type": "http",
      "url": "https://happyhorse.mcp.acedata.cloud/mcp"
    }
  }
}

README

Happy Horse MCP Server

[PyPI] [Python] [License]

Model Context Protocol server for Happy Horse AI video generation and editing through the Ace Data Cloud API.

Capabilities

Text-to-video generation
First-frame image-to-video animation
Reference-to-video generation with 1-9 subject or style images
Video editing with up to 5 reference images
720P and 1080P output
Single and batch task polling
Local stdio and hosted Streamable HTTP/SSE transports
Direct Bearer token and AceDataCloud OAuth authentication

Install

Get a token from platform.acedata.cloud.

Configure

Claude Desktop

json

{  "mcpServers": {    "happyhorse": {      "command": "uvx",      "args": ["mcp-happyhorse"],      "env": {        "ACEDATACLOUD_API_TOKEN": "your-token"      }    }  }}

{  "mcpServers": {    "happyhorse": {      "command": "uvx",      "args": ["mcp-happyhorse"],      "env": {        "ACEDATACLOUD_API_TOKEN": "your-token"      }    }  }}

Hosted MCP

json

{  "mcpServers": {    "happyhorse": {      "url": "https://happyhorse.mcp.acedata.cloud/mcp",      "headers": {        "Authorization": "Bearer your-token"      }    }  }}

{  "mcpServers": {    "happyhorse": {      "url": "https://happyhorse.mcp.acedata.cloud/mcp",      "headers": {        "Authorization": "Bearer your-token"      }    }  }}

The hosted endpoint also supports OAuth-capable MCP clients.

Tools

Tool	Purpose
`happyhorse_generate_video`	Generate a video from text
`happyhorse_generate_video_from_image`	Animate one first-frame image
`happyhorse_generate_video_from_references`	Generate from 1-9 reference images
`happyhorse_edit_video`	Edit a source video with up to 5 references
`happyhorse_get_task`	Query one task
`happyhorse_get_tasks_batch`	Query multiple tasks
`happyhorse_list_models`	List valid models for each action

Generation tools submit asynchronously when no callback_url is supplied. Keep the returned task_id, wait about 15 seconds, then call happyhorse_get_task until the response contains a final video_url or terminal error.

Models

Action	Models	Default
Text-to-video	`happyhorse-1.0-t2v`, `happyhorse-1.1-t2v`	`happyhorse-1.1-t2v`
Image-to-video	`happyhorse-1.0-i2v`, `happyhorse-1.1-i2v`	`happyhorse-1.1-i2v`
Reference-to-video	`happyhorse-1.0-r2v`, `happyhorse-1.1-r2v`	`happyhorse-1.1-r2v`
Video edit	`happyhorse-1.0-video-edit`	`happyhorse-1.0-video-edit`

Generation duration is 3-15 seconds. Supported resolutions are 720P and 1080P. Text and reference generation support 16:9, 9:16, 1:1, 4:3, and 3:4. Image-to-video follows the input image ratio. Video-edit duration follows the source video.

Example Requests

Ask your MCP client:

Generate a 720P, 9:16 video of a white horse crossing a snowy ridge at sunrise.

Animate https://example.com/horse.jpg with a slow camera push and wind moving the mane.

Edit https://example.com/source.mp4 to preserve the camera motion but apply the costume style from https://example.com/reference.jpg. Keep the original audio.

Environment

Variable	Default	Purpose
`ACEDATACLOUD_API_TOKEN`	none	API token for local stdio mode
`ACEDATACLOUD_API_BASE_URL`	`https://api.acedata.cloud`	API origin
`HAPPYHORSE_REQUEST_TIMEOUT`	`60`	HTTP request timeout in seconds
`MCP_TRANSPORT`	`stdio`	`stdio` or `http`
`MCP_SERVER_URL`	none	Public URL that enables hosted OAuth
`LOG_LEVEL`	`INFO`	Logging level

Development

Documentation

License

MIT

来源：README.md，提交 2be1ac9

工具

工具元数据尚未被收录。

版本历史

v2026.10.4.1最新Oct 4, 2026