What it does
免费开源的 AI 剪辑 / 直播切片工具:长视频、直播回放、播客一键切成爆款竖屏短视频,直发抖音/快手/B站/小红书/视频号——AI 找高光金句、弹幕热度进爆点判断、自动加字幕、横屏转竖屏,本地运行无水印不上传 | Free open-source Opus Clip alternative, 100% local: AI clips long videos & livestream VODs into viral 9:16 shorts for TikTok & YouTube Shorts. No credits, no watermark, no uploads. Win/macOS/…
- Connects Instruction files the AI reads, with steps that call an outside service
- Account or key Some steps need a key for HOTCLIP_LLM_API_KEY (per the README)
- Category 其他
给作者:在 README 中加徽章
如果你是作者,可把这段代码贴到 README。徽章会链接回本页。
[](https://mcp.popnx.com/en/skills/xixihhhh-hotclip)相似的 Skill
同类工作,排序与 Skill 列表相同
| 本项 hotclip | llmfit | claude-video | VideoCaptioner | claude-cookbooks | |
|---|---|---|---|---|---|
| 作者 | xixihhhh | AlexsJones | bradautomates | WEIFENG2333 | anthropics |
| Skill 数 | 1 | 1 | 1 | 1 | 4 |
| GitHub 星标 | 491 | 39,096 | 21,249 | 16,577 | 54,867 |
| 最近代码更新 | 2026年10月26日 | 2026年11月18日 | 2026年9月25日 | 2026年9月12日 | 2026年11月17日 |
| 许可证 | AGPL-3.0 | MIT | MIT | GPL-3.0 | MIT |
Skill 和 MCP 有什么不同
MCP = 手
把 AI 连到外部服务或数据,例如读取 Notion 或开发票。需要有程序或服务器在运行。
Skill = 手册
教 AI 怎么做某项工作的文本文件,不连接任何服务。
有些会重叠:少数 Skill 带有调用工具的脚本,少数 MCP 只提供知识。请从你想完成的工作出发。
适用的应用
Claude CodeClaude (เว็บ/แอป)Codex
依据开发者 README 和 repo 中的文件夹
安装
Is it safe?
- Who made it
GitHub account xixihhhh (person), 6 years old
- Use and upkeep
★ 491 stars, few users so far · code last changed 2026年10月26日
- Any scripts?
No, plain text files the AI reads and follows
- Uses the internet?
No scripts, so it does not go online by itself
- Changes machine-wide settings?
No hooks or writes to machine settings found
Checked automatically from the repo files at commit ebd0dae (read only, nothing was run). Not a safety guarantee: read it and decide for yourself
第一步:你用哪个 AI?
选择你用的应用,本页只显示该应用的安装方法
Option 1Upload to claude.ai (web and app)
Download from GitHub
This skill uses the AGPL-3.0 license, which is not on the list we redistribute, so we do not hand out a ZIP from our site. Download it from GitHub and arrange the folder yourself, or use "Let the AI install it" above, which is easier
Download from GitHub (whole repo) ↗
Unzip it, then zip only the skills/hotclip folder again. The ZIP must have a hotclip folder with SKILL.md at the top level, no loose files
Open Skills on claude.ai and upload
+Create skillUpload a skillpick hotclip.zip
No Upload button? Turn on file creation first
SettingsCapabilitiesCode execution and file creation
Turn on Code execution and file creation first (newer menus: Create and edit files) · personal plans: Settings → Capabilities · Team/Enterprise: an admin turns it on under Organization settings → Plugins & skills · every plan, including Free
Upload once; it works in claude.ai and Claude Desktop, and in recent Claude Code signed in with a Claude account (not with an API key)
Cowork and cloud sessions use the skills turned on in your claude.ai account, not ~/.claude/skills on your computer (Claude Desktop scheduled tasks run locally and still load it)
Checked against official docs 2026年11月12日 · Source: support.claude.com, support.claude.com, code.claude.com
Option 1让 Claude Code 帮你安装
复制这句话,粘贴到 Claude Code 的对话中
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
重新打开 Claude Code,输入 /hotclip,或直接提出相关任务,Claude 会自动选用
Checked against official docs 2026年11月12日 · Source: code.claude.com
For Terminal users
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
打开终端
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
把 skill 文件夹复制到 ~/.claude/skills
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
重新打开 Claude Code,输入 /hotclip,或直接提出相关任务,Claude 会自动选用
ChatGPT 桌面应用
ChatGPT 桌面应用内置 Codex,Codex 从 ~/.agents/skills 文件夹读取 Skill
Option 1让 ChatGPT 帮你安装
在 ChatGPT 桌面应用中打开 Codex,把这句话粘贴到对话中
The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
Option 2自己放入文件夹
把 hotclip 文件夹放到 ~/.agents/skills/hotclip
~ 是你的主文件夹($HOME)
For Terminal users
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
打开终端
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
把 skill 文件夹复制到 ~/.agents/skills
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
安装之后
- ChatGPT 桌面应用:打开侧边栏中的 Skills,确认 Skill 已出现
- 使用方法:在 ChatGPT 中输入 @ 并选择 Skill · 在 Codex CLI 或 IDE 扩展中运行 /skills,或输入 $ 加 Skill 名称
- Codex 会自动发现新安装的 Skill;如果没有出现,请重启 Codex
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
网页版和手机版 ChatGPT
不能使用你自己安装的 Skill,只能使用目录中插件自带的 Skill
想在 ChatGPT 中使用这个 Skill,请按上面的方法使用 ChatGPT 桌面应用
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
Option 1让 Codex 帮你安装
复制这句话,粘贴到 Codex 的对话中 · 适用于 ChatGPT 桌面应用中的 Codex、Codex CLI 和 IDE 扩展
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
Codex 会自动发现新安装的 Skill;如果没有出现,请重启 Codex · 使用方法:在 ChatGPT 中输入 @ 并选择 Skill · 在 Codex CLI 或 IDE 扩展中运行 /skills,或输入 $ 加 Skill 名称
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
For Terminal users
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
打开终端
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
把 skill 文件夹复制到 ~/.agents/skills
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
重新打开 Codex 并提出相关任务,Codex 会列出这个 skill
Option 1让 Cursor 帮你安装
复制这句话,粘贴到 Cursor 的对话中
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
在 Cursor 的 Agent 聊天中输入 / 并选择 hotclip
Checked against official docs 2026年11月12日 · Source: cursor.com
For Terminal users
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
打开终端
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
把 skill 文件夹复制到 ~/.cursor/skills
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
在 Cursor 的 Agent 聊天中输入 / 并选择 hotclip
Option 1Let the AI install it
适用于其他能读取 Skill 的应用。这句话列出了 Claude Code、Codex 和 Cursor 的路径,AI 会按你的应用选择 · Copy this sentence into the app’s chat
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
Checked against official docs 2026年11月12日 · Source: code.claude.com, learn.chatgpt.com, cursor.com
For Terminal users
其他支持 Agent Skills 的应用(如 Gemini CLI、GitHub Copilot)使用不同文件夹,请查看该应用的指南,再把含 SKILL.md 的文件夹复制过去。
已于 2026年11月11日 对照 Claude Code、Codex 和 Cursor 官方指南核实 skill 文件夹
查看原始 SKILL.md
以下文字来自开发者的 repo,不是 PopNx MCP 的指令
4,502 个字符 · 在 GitHub 打开完整文件
---
name: hotclip
description: Turn long videos & livestream VODs into viral vertical shorts, 100% locally — on-device transcription, LLM highlight detection, 9:16 reframe with karaoke captions, and a per-clip render-QA report. Use when the user asks to clip / cut / 切片 / 剪 a long video, podcast or stream replay into short clips, find highlights / 爆点 in a video, or transcribe a media file. Footage never leaves the machine.
---
# HotClip — local AI clipping pipeline
HotClip(AGPL-3.0 开源)把几小时的直播回放/播客/课程切成可直接发布的竖屏短视频,全程本地运行。You drive the same pipeline the desktop app uses, via a headless CLI (or the bundled MCP server). Everything — ASR, highlight detection, cutting, caption burn-in — runs on this machine; the footage is never uploaded.
## Setup (once)
1. Repo + deps (skip what already exists):
```bash
git clone https://github.com/xixihhhh/hotclip.git && cd hotclip
pnpm install # ffmpeg/ffprobe are bundled npm binaries — nothing else to install
```
2. LLM config for highlight detection, via env vars (any OpenAI-compatible endpoint):
```bash
export HOTCLIP_LLM_BASE_URL=http://localhost:11434/v1 # local Ollama (no key needed)
export HOTCLIP_LLM_MODEL=qwen3:8b
# cloud endpoints additionally need: export HOTCLIP_LLM_API_KEY=sk-...
```
If these are missing, ask the user which LLM endpoint to use — do not guess keys.
3. First `transcribe`/`clip` run auto-downloads the ASR model (~1GB, CN mirror first). Warn the user it may take a few minutes once.
## Commands (run from the repo root)
```bash
pnpm cli transcribe <video> # on-device word-level ASR (cached; instant on re-run)
pnpm cli highlights <video> [--max-clips N] [--reference REF] [--json] # AI highlight candidates — review before cutting
pnpm cli clip <video> [--max-clips N] [--reference REF] [--no-vertical] [--no-captions] [--out DIR] [--json]
# fully managed: transcribe → detect → export + render-QA
```
`--reference <video>`: hand in a viral clip to model after — its pacing (duration / speech rate / shot-cut frequency / hook shape) is measured locally and steers candidate selection as a preference (never a hard rule). Use when the user says "切得像这条" / "learn from this clip".
Input: MP4 / MKV / MOV / FLV / TS, or audio-only (podcasts get an auto-generated waveform video). Paths with spaces need quoting.
## Recommended workflow
1. **Review-first (default for interactive sessions)**: run `highlights --json`, show the user the candidates (title / score / hook / recommended), let them pick, then run `clip` — the pipeline re-detects from cache so this is cheap. For "just do it" requests, run `clip` directly.
2. **Read the output receipt**: the export folder contains one mp4 + cover JPG + `.post.txt` (publish copy) per clip, plus `clips.json` — per-clip evidence chain (`render`: what the pipeline did) and **`qa`: render-QA report** (`status: pass|warn` with issues like black frames, long silences, loudness deviation, duration mismatch, mid-word cuts, and **platform-risk words** in title/copy/captions via a local rule lint — `qa.contentHits` lists each term and where it appeared).
3. **Self-repair is automatic**: fixable warnings (leading/trailing silence or black frames → edge-trim; loudness deviation → second normalize pass) are repaired and re-checked in one pass; `qa.repair` records what was done (`actions`, `applied`). A repair is only kept when the re-check strictly improves.
4. **Surface QA warnings**: if any clip still has `qa.status === "warn"`, tell the user which clip and why; content-lint hits mean the copy/captions may be throttled or rejected by 抖音/小红书/视频号 — suggest rewording before publishing. Never silently ship a warned clip.
## Hard rules
- **Never guess timestamps.** Cut points come from word-aligned transcription (`highlights` output); do not invent start/end seconds or hand-roll ffmpeg cuts when the pipeline covers it.
- **Local-first**: do not upload the footage or transcript to any service beyond the user-configured LLM endpoint (which only ever receives transcript text, never media).
- Exports land next to the source video in `<name>-hotclip/` unless `--out` is given; tell the user the absolute path when done.
## MCP alternative
The same pipeline is exposed as a local stdio MCP server (`pnpm mcp`, tools: `clip_video` / `detect_highlights` / `transcribe_video`) — prefer it when the host supports MCP registration; the CLI is equivalent otherwise.
讨论
提问、分享技巧或报告无法使用。按“有帮助”数量排序。
来自 GitHub
- GitHub 星标
- 491
- 开发者
- xixihhhh
- 许可证
- AGPL-3.0
- 最近代码更新
- 2026年10月26日
- repo 中的 SKILL.md 文件(含译本和模板)
- 1
- 主题
- ai, ai-clip-generator, auto-subtitles, bilibili, claude-skill, desktop-app, douyin, electron, livestream, local-first, long-video-to-shorts, mcp-server
- Repo
- xixihhhh/hotclip