SeaArt Video Generator
This skill helps you generate AI videos via the SeaArt platform. You can create videos from text prompts (Text-to-Video) or animate existing images (Image-to-Video).
Prerequisites: SeaArt Token
To use this skill, you need a SeaArt authentication token. This token is passed in the T cookie.
How to get your token:
- 1. Log into your account on seaart.ai in your browser.
- Open Developer Tools (F12 or Right Click -> Inspect).
- Go to the "Application" tab (or "Storage" in Firefox).
- Expand "Cookies" and select
https://www.seaart.ai. - Find the cookie named
T. - Copy its value (it will look like a long string:
eyJhbGci...).
How to store your token:
To avoid passing the token every time, store it as an environment variable (
SEAART_TOKEN).
Run the following command in your terminal (never paste the token into chat):
CODEBLOCK0
Security note: Never share your token in chat messages or logs. The token will be persisted in your local agent config by /update-config — ensure only you have access to that config file.
Supported Models and Parameters
We currently support the following models for video generation.
Default Model: SeaArt Sono Cast (if user doesn't specify, use this one)
| Model Name | INLINECODE6 | INLINECODE7 | Notes |
|---|
| SeaArt Sono Cast (Default) | INLINECODE8 | INLINECODE9 | Great overall quality. |
| Vidu Q3 Turbo |
d6l3adte878c73dr6a3g |
26ca53e1-e6f2-44e4-8227-13e38bb73945 | |
|
Kling 3.0 |
d62lo1te878c73f8eua0 |
93636e78-3d55-4416-92e3-d7d3b89f023d | |
|
Wan 2.6 |
d4t6pkle878c73cpsif0 |
5856d251-5b8f-4704-970e-e301982eb532 | |
Supported Aspect Ratios (Text-to-Video only)
- -
16:9 (Landscape) - INLINECODE17 (Portrait) - Default
- INLINECODE18 (Square)
- INLINECODE19
- INLINECODE20
How to use this skill
When a user asks to generate a video, follow these steps:
1. Check for the Token
First, check if the
SEAART_TOKEN environment variable is set. You can do this by running a simple bash command:
echo $SEAART_TOKEN.
If it's empty, ask the user to provide their token. Once they provide it, you should guide them to set it as an environment variable in their specific environment (OpenClaw, Claude Code, or terminal) as described above. Wait for them to do this before proceeding.
2. Gather Requirements
If the token is available, ensure you have:
- - The prompt (what should happen in the video)
- The type (Text-to-Video or Image-to-Video)
- If Image-to-Video, the image URL
- The desired model (default to SeaArt Sono Cast)
- If Text-to-Video, the desired aspect ratio (default to 9:16)
3. Generate the Video
Use the included python script to start the generation and poll for completion.
CODEBLOCK1
The script will handle making the request and polling the status until the video is complete, then return the final video URL.
Example Usage
User: "Can you make a video of a cat playing with a ball of yarn using Kling 3.0?"
Claude:
- 1. Checks for
$SEAART_TOKEN. Let's assume it's set. - Runs the generation script:
python3 ~/.claude/skills/seaart-video/scripts/generate.py --type t2v --prompt "A cat playing with a ball of yarn" --model kling3.0
- 3. Presents the final video URL to the user.
External Endpoints
| URL | Method | Data Sent |
|---|
| INLINECODE24 | POST | Prompt, model ID, aspect ratio, generation params |
| INLINECODE25 |
POST | Prompt, image URL, model ID, generation params |
|
https://www.seaart.ai/api/v1/task/batch-progress | POST | Task ID (for polling status) |
All requests are authenticated via your SEAART_TOKEN cookie, which never leaves your machine in plaintext — it is only sent to *.seaart.ai as an HTTP cookie header.
Security & Privacy
- - What leaves your machine: Your text prompt, any image URLs you provide, and your
SEAART_TOKEN are sent to seaart.ai servers as an HTTP cookie header on every API request. - Token storage:
SEAART_TOKEN is persisted locally in your agent config file by /update-config. Ensure only you have access to that config. It is not logged or transmitted by this skill beyond the API calls to seaart.ai listed above. - No additional data persistence: This skill does not write any other files to disk. Generated video URLs are returned inline.
- Input handling: All inputs are passed as structured JSON fields via the
requests library — no shell interpolation is used.
Trust Statement
By using this skill, your prompt and any provided image URLs are sent to seaart.ai. Only install and use this skill if you trust SeaArt with your creative content. This skill is not affiliated with or endorsed by SeaArt.
SeaArt 视频生成器
此技能可帮助您通过 SeaArt 平台生成 AI 视频。您可以根据文本提示创建视频(文本转视频),或为现有图像添加动画效果(图像转视频)。
前提条件:SeaArt 令牌
要使用此技能,您需要一个 SeaArt 身份验证令牌。该令牌通过 T cookie 传递。
如何获取令牌:
- 1. 在浏览器中登录您的 seaart.ai 账户。
- 打开开发者工具(F12 或右键点击 -> 检查)。
- 转到应用程序选项卡(Firefox 中为存储)。
- 展开Cookies并选择 https://www.seaart.ai。
- 找到名为 T 的 cookie。
- 复制其值(看起来像一长串字符:eyJhbGci...)。
如何存储令牌:
为避免每次传递令牌,请将其存储为环境变量(SEAART_TOKEN)。
在终端中运行以下命令(切勿将令牌粘贴到聊天中):
/update-config set SEAARTTOKEN=yourtokenvaluehere
安全提示: 切勿在聊天消息或日志中分享您的令牌。令牌将通过 /update-config 持久保存在您的本地代理配置中 — 请确保只有您能访问该配置文件。
支持的模型和参数
我们目前支持以下视频生成模型。
默认模型: SeaArt Sono Cast(如果用户未指定,则使用此模型)
| 模型名称 | modelno | modelver_no | 备注 |
|---|
| SeaArt Sono Cast(默认) | d4t763te878c73csvcf0 | 21a5fc20-68dd-4d4b-9afd-7bfe2e8f5166 | 整体质量优秀。 |
| Vidu Q3 Turbo |
d6l3adte878c73dr6a3g | 26ca53e1-e6f2-44e4-8227-13e38bb73945 | |
|
Kling 3.0 | d62lo1te878c73f8eua0 | 93636e78-3d55-4416-92e3-d7d3b89f023d | |
|
Wan 2.6 | d4t6pkle878c73cpsif0 | 5856d251-5b8f-4704-970e-e301982eb532 | |
支持的宽高比(仅限文本转视频)
- - 16:9(横屏)
- 9:16(竖屏)- 默认
- 1:1(正方形)
- 4:3
- 3:4
如何使用此技能
当用户要求生成视频时,请按照以下步骤操作:
1. 检查令牌
首先,检查是否设置了 SEAART
TOKEN 环境变量。您可以通过运行一个简单的 bash 命令来执行此操作:echo $SEAARTTOKEN。
如果为空,请要求用户提供其令牌。一旦他们提供,您应指导他们按照上述说明将其设置为特定环境(OpenClaw、Claude Code 或终端)中的环境变量。等待他们完成此操作后再继续。
2. 收集需求
如果令牌可用,请确保您拥有:
- - 提示词(视频中应发生什么)
- 类型(文本转视频或图像转视频)
- 如果是图像转视频,需要图像 URL
- 所需的模型(默认为 SeaArt Sono Cast)
- 如果是文本转视频,所需的宽高比(默认为 9:16)
3. 生成视频
使用附带的 Python 脚本启动生成并轮询完成状态。
bash
文本转视频
python3 ~/.claude/skills/seaart-video/scripts/generate.py \
--type t2v \
--prompt 一个哭泣的孩子 \
--model seaart-sono-cast \
--aspect-ratio 9:16
图像转视频
python3 ~/.claude/skills/seaart-video/scripts/generate.py \
--type i2v \
--prompt 他结婚了 \
--image-url https://image.cdn2.seaart.me/... \
--model seaart-sono-cast
该脚本将处理请求并轮询状态,直到视频完成,然后返回最终的视频 URL。
使用示例
用户:你能用 Kling 3.0 制作一个猫玩毛线球的视频吗?
Claude:
- 1. 检查 $SEAART_TOKEN。假设已设置。
- 运行生成脚本:
bash
python3 ~/.claude/skills/seaart-video/scripts/generate.py --type t2v --prompt 一只猫在玩毛线球 --model kling3.0
- 3. 向用户呈现最终的视频 URL。
外部端点
| URL | 方法 | 发送的数据 |
|---|
| https://www.seaart.ai/api/v1/task/v2/video/text-to-video | POST | 提示词、模型 ID、宽高比、生成参数 |
| https://www.seaart.ai/api/v1/task/v2/video/img-to-video |
POST | 提示词、图像 URL、模型 ID、生成参数 |
| https://www.seaart.ai/api/v1/task/batch-progress | POST | 任务 ID(用于轮询状态) |
所有请求均通过您的 SEAART_TOKEN cookie 进行身份验证,该令牌不会以明文形式离开您的机器 — 它仅作为 HTTP cookie 标头发送到 *.seaart.ai。
安全与隐私
- - 离开您机器的内容:您的文本提示词、您提供的任何图像 URL 以及您的 SEAARTTOKEN 会在每次 API 请求中作为 HTTP cookie 标头发送到 seaart.ai 服务器。
- 令牌存储:SEAARTTOKEN 通过 /update-config 持久保存在您的本地代理配置文件中。请确保只有您能访问该配置。除上述对 seaart.ai 的 API 调用外,此技能不会记录或传输该令牌。
- 无额外数据持久化:此技能不会将任何其他文件写入磁盘。生成的视频 URL 会直接返回。
- 输入处理:所有输入均通过 requests 库以结构化 JSON 字段形式传递 — 不使用 shell 插值。
信任声明
使用此技能即表示您的提示词和任何提供的图像 URL 将被发送到 seaart.ai。仅当您信任 SeaArt 处理您的创意内容时才安装和使用此技能。此技能与 SeaArt 无关联,也未获得其认可。