← Back to Skills Marketplace
don068589

Douyin Orchestrator

by Don Li · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
133
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install douyin-orchestrator
Description
抖音视频转录工作流编排。从抖音链接下载视频、转录语音、分析内容,输出结构化成果。编排调用 douyin-fetcher、douyin-transcriber、douyin-analyzer 三个独立模块。
README (SKILL.md)

Douyin Transcribe - 工作流编排

** 完整的抖音视频转录工作流,按顺序调用三个独立模块:

  1. douyin-fetcher - 获取视频
  2. douyin-transcriber - 转录音频
  3. douyin-analyzer - 分析内容

架构

用户发送抖音链接
    ↓
[Step 0] 判断输入类型
    ↓
[Step 1] fetcher: 解析链接 → 下载视频
    ↓
[Step 2] transcriber: 提取音频 → 转录文字
    ↓
[Step 3] analyzer: 清理文本 → 分析要点
    ↓
[Step 4] 保存成果 → 更新索引

完整工作流

Step 0: 判断输入类型

输入类型 处理方式
视频链接(/video/ 完整工作流
图文链接(/note/ 浏览器 snapshot 提取文字
本地视频文件 从 Step 2 开始
文本内容 从 Step 3 开始

Step 1: 下载视频

调用模块douyin-fetcher

详见 ~/.openclaw/skills/douyin-fetcher/SKILL.md

⚠️ 注意:Step 1 的 browser act 调用必须用 request={"kind": "evaluate", "fn": "..."} 嵌套格式。

输出:视频流 + 音频流(DASH)或完整视频文件


Step 2: 转录音频

调用模块douyin-transcriber

详见 ~/.openclaw/skills/douyin-transcriber/SKILL.md

⚠️ 注意:如果 Step 1 下载了 DASH 分离流,直接用 audio.mp4 转 WAV 转录,不需要先合并

输出:转录文本


Step 3: 分析内容

调用模块douyin-analyzer

详见 ~/.openclaw/skills/douyin-analyzer/SKILL.md

做法:转录后由 agent 直接分析(语义分段、修正错误、提取要点)


Step 4: 保存成果

转录稿(/path/to/knowledge ranscripts\)

文件 说明
{主题}-完整转录.md 包含标题、作者、链接、转录时间、完整内容

视频文件(/path/to/videos\)

文件 说明
{主题}.mp4 原始视频(可选保存)

更新索引

保存文件后,更新:

  • /path/to/knowledge ranscripts\data_structure.md
  • /path/to/knowledge\data_structure.md(主索引)
  • /path/to/videos\data_structure.md(如保存视频)

部分工作流

用户请求 执行步骤
"帮我转录这个抖音视频" Step 1 → 2 → 3 → 4
"只下载这个视频,不转录" Step 1 → 保存到 /path/to/videos\
"我已经有视频文件了,帮我转录" Step 2 → 3 → 4
"帮我分析这段文本" Step 3 → 4

异常处理汇总

阶段 问题 处理
Step 1 短链解析失败 检查链接完整性,去掉分享文案干扰字符
Step 1 JS 返回空数组 等待 15-20 秒后重试,或扩大搜索范围
Step 1 下载 403 URL 过期,重新获取
Step 1 浏览器不可用 提示用户手动下载
Step 2 ffmpeg 未安装 winget install Gyan.FFmpeg
Step 2 Whisper 服务未运行 docker start whisper-asr
Step 2 转录超时 长视频需要更多时间,10分钟视频约需 15-20 分钟
Step 2 转录乱码 语言识别问题,可后续手动修正
Step 4 索引更新失败 手动更新 data_structure.md

图文笔记处理

图文链接(/note/)无需下载,直接提取:

1. browser(action='open', profile='openclaw', url='图文链接')
2. browser(action='snapshot')
3. 从 snapshot 提取文字内容
4. 保存到 /path/to/knowledge	ranscripts\

前置条件

工具 用途 检查命令
curl 下载文件 curl --version
ffmpeg 视频合并/格式转换 ffmpeg -version
Docker Whisper 转录服务(内含 ffmpeg 7.1) docker ps --filter "name=whisper-asr"
浏览器 抓取视频 openclaw profile 可用

说明:宿主机 ffmpeg 8.1 用于视频合并等任务,Whisper ASR 容器内 ffmpeg 7.1 用于转录处理,两者并存。


已知限制

  1. 视频 URL 有时效性 — 获取后立即下载
  2. 长视频转录耗时 — 10分钟视频约需 15-20 分钟
  3. 需要浏览器 — openclaw profile 必须可用
  4. 图文笔记 — 部分需要登录才能查看完整内容
  5. 部分视频无分离流 — 直接是完整 MP4,无需合并
Usage Guidance
This skill is an orchestration recipe (no code bundled) and appears coherent for video→transcript→analysis tasks, but it relies on other skills (douyin-fetcher, douyin-transcriber, douyin-analyzer) which are not included — review those modules' SKILL.md/code before installation. Be cautious about the browser evaluate step: executing arbitrary JavaScript in a page context can access cookies, auth tokens, or perform network requests, so only run this with a browser/profile you trust. Confirm the whisper-asr Docker image and any ffmpeg installation come from trusted sources. Verify and replace placeholder paths (/path/to/...) with safe directories and ensure you are comfortable the skill will write transcripts/videos and update index files. If you need higher assurance, request the source for the referenced submodules or run the workflow in an isolated environment (container/VM) first.
Capability Assessment
Purpose & Capability
Name/description match the instructions: it orchestrates three submodules (fetcher, transcriber, analyzer), declares no unrelated env vars or installs, and lists reasonable prerequisites (browser, ffmpeg, Docker). The files and pointers are consistent with an orchestrator role.
Instruction Scope
SKILL.md limits itself to orchestration steps, file locations, and calling other skills; it instructs the agent to run browser actions (including evaluate with arbitrary JS) and to read/write user paths (transcripts, videos, index files). These actions are coherent for a fetch/transcribe/analyze workflow but carry intrinsic risk (browser-evaluated JS can access page context, cookies, or cause network requests). The skill itself does not instruct reading unrelated system files or environment variables.
Install Mechanism
No install spec or external downloads — instruction-only. This minimizes direct install risk; nothing is written to disk by the skill package itself.
Credentials
No environment variables, credentials, or config paths are requested. The skill instead expects local tools/containers (ffmpeg, Docker/whisper) which are proportionate to media processing.
Persistence & Privilege
always:false and normal autonomous invocation. The skill instructs saving files and updating index markdowns under user paths (placeholders shown). This is expected for a workflow that persists transcripts, but users should confirm the exact filesystem paths and permissions before use.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install douyin-orchestrator
  3. After installation, invoke the skill by name or use /douyin-orchestrator
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release - Workflow orchestration module
Metadata
Slug douyin-orchestrator
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Douyin Orchestrator?

抖音视频转录工作流编排。从抖音链接下载视频、转录语音、分析内容,输出结构化成果。编排调用 douyin-fetcher、douyin-transcriber、douyin-analyzer 三个独立模块。 It is an AI Agent Skill for Claude Code / OpenClaw, with 133 downloads so far.

How do I install Douyin Orchestrator?

Run "/install douyin-orchestrator" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Douyin Orchestrator free?

Yes, Douyin Orchestrator is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Douyin Orchestrator support?

Douyin Orchestrator is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Douyin Orchestrator?

It is built and maintained by Don Li (@don068589); the current version is v1.0.0.

💬 Comments