douyin-video-summary — Claude Code skill
Pulls audio from Douyin videos, transcribes it locally with whisper.cpp, and turns the transcript into a structured summary.
npx skills add liu-wei-ai/douyin-video-summary --skill douyin-video-summary -a claude-codeWhat douyin-video-summary does
The skill responds to v.douyin.com and douyin.com/video/ links. It resolves the video ID, opens the page in a browser, and captures audio requests because direct Douyin CDN downloads through yt-dlp or aria2c return 403. The captured file is fetched with curl and a Referer header.
ffmpeg converts the audio to mono WAV at 16000 Hz. whisper.cpp then creates a transcript, which the agent turns into a title, key points, and a short takeaway. The workflow will not run without ffmpeg, whisper.cpp, and the downloaded ggml-small.bin model.
Setup requires whisper-cpp, ffmpeg, and a GGML model. An optional Feishu sync can append the summary to a Feishu document through the Feishu Open API. Metal acceleration is automatic on Apple Silicon; the medium model may run out of memory on an 8GB machine.
What it helps with
- extract audio from a Douyin link
- transcribe Chinese speech on the local machine
- turn a transcript into structured points
- append a summary to a Feishu document
Example prompts for the agent
- Summarize the video at
https://v.douyin.com/xxxxx/and list its main points. - Open
https://www.douyin.com/video/7604713801732365681, transcribe the speech, and write a structured summary. - Sync this Douyin summary to the configured Feishu document.
How to install douyin-video-summary
1Skills CLI (Vercel)
npx skills add liu-wei-ai/douyin-video-summary --skill douyin-video-summary -a claude-code2Manually: put the files in place
~/.claude/skills/douyin-video-summary/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/douyin-video-summary/
1Skills CLI (Vercel)
npx skills add liu-wei-ai/douyin-video-summary --skill douyin-video-summary -a codex2Manually: put the files in place
.agents/skills/douyin-video-summary/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add liu-wei-ai/douyin-video-summary --skill douyin-video-summary -a cursor2Manually: put the files in place
.cursor/skills/douyin-video-summary/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install liu-wei-ai/douyin-video-summary douyin-video-summary2Manually: put the files in place
.github/skills/douyin-video-summary/SKILL.md1Skills CLI (Vercel)
npx skills add liu-wei-ai/douyin-video-summary --skill douyin-video-summary -a gemini-cli2Platform CLI
gemini skills install https://github.com/liu-wei-ai/douyin-video-summaryManually: put the files in place
.devin/skills/douyin-video-summary/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add liu-wei-ai/douyin-video-summary --skill douyin-video-summary -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 douyin-video-summary
How do I install douyin-video-summary in Claude Code?
Run npx skills add liu-wei-ai/douyin-video-summary --skill douyin-video-summary -a claude-code, then check ~/.claude/skills/douyin-video-summary/SKILL.md.
What does douyin-video-summary do?
It captures audio from a Douyin video, creates a local transcript with whisper.cpp, and produces a structured summary. The result can also be synced to Feishu.
What does the skill require?
Install whisper-cpp and ffmpeg, then download the ggml-small.bin GGML model. Audio capture uses a browser and curl; aria2c receives a 403 response from Douyin CDN URLs.
