← Back to Skills Marketplace
sunshinehu

550W视频去字幕

by SunshineHu · GitHub ↗ · v1.0.1 · MIT-0
cross-platform ✓ Security Clean
93
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install 550w-ai-subtitle-remover
Description
上传视频或提供 URL,调用 550W AI 去字幕 API 擦除硬字幕/水印/台标/Logo,返回处理后视频
README (SKILL.md)

AI 视频去字幕

When to Use

  • 用户需要去除视频中的硬字幕、水印、台标或 Logo
  • 用户提供了本地视频文件或视频 URL,希望得到去字幕后的干净视频
  • 用户需要查询去字幕任务的处理进度或历史记录
  • 用户需要查询账户积分余额或预估处理费用

凭证配置

首次使用前需要配置凭证:

支持的 Actions

Action 说明
uploadVideo 上传本地视频文件,获取云端 URL
submitTask 提交去字幕任务,指定字幕区域
taskDetail 查询任务处理状态和结果
taskList 分页查询历史任务列表
queryCredits 查询账户积分余额,支持消耗预估
workflow 端到端工作流:上传→提交→轮询→返回结果

Workflow

重要:调用规则

调用 workflow action 时,只传 filevideoUrl 一个参数即可。禁止传递 width、height、duration、x1、y1、x2、y2 参数。 Skill 内部会自动获取视频元信息并使用全屏去字幕模式。不要使用 ffprobe 或其他工具预先探测视频信息。

正确调用示例:

{ "action": "workflow", "params": { "videoUrl": "https://example.com/video.mp4" } }

或:

{ "action": "workflow", "params": { "file": \x3C视频文件对象> } }

错误调用示例(不要这样做):

{ "action": "workflow", "params": { "videoUrl": "...", "width": 1920, "height": 1080, "duration": 60, "x1": 0, "y1": 0, "x2": 0, "y2": 0 } }

执行流程

  1. 用户提供视频文件或视频 URL。
  2. 直接调用 workflow action,仅传 filevideoUrl
  3. Skill 自动获取 width/height/duration,坐标默认全屏模式。
  4. Skill 自动完成上传(如需)、提交任务、轮询状态,最终返回去字幕后的视频下载地址。
  5. 如果任务在 10 分钟内未完成,返回 taskId 供后续手动查询。

分步模式

  1. 调用 uploadVideo 上传视频文件,获取 videoUrl。
  2. 调用 submitTask 提交去字幕任务,指定 videoUrl、宽高、时长和字幕区域坐标。
  3. 调用 taskDetail 轮询任务状态,直到 status 为 success 或 failed。
  4. 成功时从 resultUrl 下载去字幕后的视频。

参数约束

  • 视频格式: 仅支持 mp4、mov
  • 文件大小: 最大 1GB
  • 分辨率: width × height ≤ 2,073,600 像素
  • 时长: 1~600 秒
  • 宽高: 1~10000 像素
  • 字幕区域: 全屏模式设 x1=y1=x2=y2=0;非全屏模式需满足 x2>x1, y2>y1, x2≤width, y2≤height
  • videoUrl: 必须以 http:// 或 https:// 开头,长度 ≤2048

积分计费

  • 720p 及以下(≤921,600 像素): ⌈时长(秒) × 1.3⌉ 积分
  • 超过 720p(>921,600 像素): ⌈时长(秒) × 1.6⌉ 积分
  • 任务失败时积分自动退还

错误处理

  • code=-100: 鉴权失败,检查 userNo 和 apiKey 配置
  • code=-200: 参数不合法,根据 message 修正参数
  • code=-300: 业务拒绝(积分不足、任务不存在等)
  • code=-500: 服务异常,建议 30 秒后重试

注意事项

  • 相同 videoUrl 重复提交会被视为独立任务并独立计费
  • workflow 模式最多轮询 10 分钟(20 次,每次间隔 30 秒)
  • 连续 3 次轮询失败时 workflow 会提前终止并返回 taskId
  • 查询接口超时 10 秒,上传接口超时 180 秒,提交接口超时 150 秒

回复用户规范

任务完成后,必须将结果告知用户:

成功时回复:

  • 告知用户去字幕已完成
  • 提供去字幕后的视频下载链接(resultUrl)
  • 可选:告知消耗积分数

失败时回复:

  • 告知用户任务失败
  • 提供失败原因(failReason)
  • 告知积分已自动退还

超时时回复:

  • 告知用户任务仍在处理中
  • 提供 taskId 供后续查询

禁止回复 NO_REPLY。任何情况下都必须向用户展示处理结果。

Usage Guidance
Install only if you trust the 550W service and the skill publisher. Treat uploaded videos as shared with the provider, protect SUBTITLE_REMOVER_USER_NO and SUBTITLE_REMOVER_API_KEY, consider checking credits before running the workflow, and avoid creating a local .credentials.json file unless you can secure it.
Capability Analysis
Type: OpenClaw Skill Name: 550w-ai-subtitle-remover Version: 1.0.1 The skill is a legitimate integration for the 550W AI subtitle removal service, providing actions to upload videos, submit processing tasks, and poll for results. The implementation in 'dist/api-client.js' and 'dist/workflow-engine.js' uses standard libraries (axios, form-data) to communicate exclusively with the service's official domains (550wai.cn). Credential management is handled through environment variables or local storage, and the instructions in 'SKILL.md' are focused on operational constraints rather than malicious prompt injection.
Capability Tags
requires-sensitive-credentials
Capability Assessment
Purpose & Capability
The advertised purpose matches the code: upload or submit a video URL, create a subtitle-removal task, poll for status, and return a result URL. This necessarily sends selected media/URLs to the 550W service and can consume account credits.
Instruction Scope
Instructions are mostly scoped to the subtitle-removal workflow. One noteworthy detail is that URL workflows rely on internal/provider metadata handling rather than agent-side probing, so users should be mindful of cost and duration assumptions.
Install Mechanism
There is no install script or shell execution shown. The package uses ordinary Node dependencies, but they are specified with semver ranges and no lockfile is included in the artifacts.
Credentials
The required userNo/apiKey credentials are proportionate for calling the stated 550W API, and the code sends them only to the configured 550W base URL.
Persistence & Privilege
The skill prefers environment variables but also contains a set method that can persist credentials to a local .credentials.json file. No background autostart, self-propagation, or hidden persistence is shown.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install 550w-ai-subtitle-remover
  3. After installation, invoke the skill by name or use /550w-ai-subtitle-remover
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
**重要调用规则与用户回复规范调整,简化为端到端全自动“workflow”调用方式。** - “workflow” action 现在只需传 videoUrl 或 file,禁止传 width/height/duration/坐标,Skill 会自动获取元信息并采用全屏去字幕。 - 去除了对 ffprobe 或其他工具预探测视频的建议,强调 Skill 必须内部自动处理视频属性。 - 新增了明确的调用示例(正确/错误)与参数填写要求,杜绝误用。 - 增加详细的用户回复规范,要求无论成功、失败或超时均主动向用户说明结果,禁止 NO_REPLY。 - 源码精简,仅保留关键运行依赖文件。
v1.0.0
ai-subtitle-remover 是一个视频去字幕 Skill,封装了 550W AI 平台的极速去字幕 OpenAPI。支持通过 Agent 直接调用,提供从视频上传到获取去字幕结果的完整能力闭环。 核心能力: 视频上传(uploadVideo):上传本地视频获取云端 URL,支持 mp4/mov 格式,最大 1GB 提交去字幕任务(submitTask):指定字幕区域坐标,支持全屏模式和区域模式 任务状态查询(taskDetail):实时查询处理进度,获取去字幕后的视频下载地址 历史任务列表(taskList):分页查询已提交的任务记录 积分余额查询(queryCredits):查询账户剩余积分,支持消耗预估 端到端工作流(workflow):一次调用完成上传→提交→轮询→返回结果的完整流程 特性: 完善的参数校验,避免无效请求 统一错误码体系(-100 鉴权/-200 参数/-300 业务/-500 服务异常) 首次使用自动引导凭证配置 workflow 模式支持 10 分钟自动轮询,超时返回 taskId 供后续查询 适用平台: SkillHub、ClawHub 适用 Agent: OpenClaw、Hermes 凭证要求: 需要 userNo 和 apiKey,申请地址 https://qzm.550wai.cn
Metadata
Slug 550w-ai-subtitle-remover
Version 1.0.1
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 2
Frequently Asked Questions

What is 550W视频去字幕?

上传视频或提供 URL,调用 550W AI 去字幕 API 擦除硬字幕/水印/台标/Logo,返回处理后视频. It is an AI Agent Skill for Claude Code / OpenClaw, with 93 downloads so far.

How do I install 550W视频去字幕?

Run "/install 550w-ai-subtitle-remover" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is 550W视频去字幕 free?

Yes, 550W视频去字幕 is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does 550W视频去字幕 support?

550W视频去字幕 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created 550W视频去字幕?

It is built and maintained by SunshineHu (@sunshinehu); the current version is v1.0.1.

💬 Comments