一条命令，把长视频变成多条带字幕短片

把任意长视频拆成N条带字幕的短片。整段视频只需转录一次，AI会自动挑选精彩时刻，再并行完成每条短片的裁剪与字幕制作。还可转换为适合TikTok、Reels和Shorts的竖屏（9:16）、适合Instagram的方形（1:1）或竖版（4:5），字幕位置会随画幅自动调整。

配方

运行

  • 视频
  • 6 个流水线
  • ≥ 11.6 积分

人工智能代理

终端

初次使用? 快速开始
01
Homebrew brew install pipe2-ai/tap/pipe2
↳
Go go install github.com/pipe2-ai/pipe2-cli/cmd/pipe2@latest
02
获取 API 令牌 ↗ echo "$PAT" | pipe2 auth login --token -

5 一集中的剪辑

更改结果

--input必填
asset_url默认—Source video: a YouTube or social URL, a direct media URL, a local file, or an existing pipe2 asset. Remote and local sources are uploaded automatically; use --asset <id> or --no-fetch when the asset is already available.
--reformat
enum默认Optional output aspect ratio. Leave empty to preserve the source's native aspect (default: fastest, cheapest). Set to 9:16 for TikTok/Reels/Shorts, 1:1 or 4:5 for Instagram, 16:9 for horizontal YouTube cards from a vertical source.
· 9:16 · 1:1 +2
9:161:14:516:9
--highlights-count
int默认5How many moments to pick when highlights runs (auto mode). Ignored if --clips is set.
--highlights-style
string默认Natural-language steer for the highlights picker: e.g. "the funniest moments", "the strongest arguments". Empty uses the picker's default.
--clips
string默认Optional path to a JSON file overriding the auto-picker, shaped [{"context": "...", "start_sec": 42.5, "end_sec": 78.0}, ...]. When set, the highlights step is skipped. Leave empty to let the highlights pipeline pick automatically.
--corrections
string默认Comma-separated word-boundary substitutions applied to the transcript, in the form "from=to,from=to". Use it when the recognizer mis-hears the same word the same way every time (a name, an acronym, a domain term). To preserve a phrase that contains a substring you also want to rewrite, declare the longer phrase first as a no-op ("phrase=phrase,word=replacement"): longer matches win, so the phrase is shielded before the bare-word rule fires. Case-sensitive.
--lang
string默认enISO 639-1 transcription language, or 'auto'.
--no-watermark
bool默认falseShip the output unbranded. Skips the watermark step entirely.
--parallel
int默认4Max number of clips to process in parallel.
--position
enum默认autoVertical caption position. "auto" (default) keeps text away from the main subject when possible and otherwise places it at the bottom. Set a position explicitly to override.
auto · top · middle +1
autotopmiddlebottom
--preset
enum默认serif-editorialCaption styling preset for every clip.
tiktok-bold-yellow · minimal-white · subtle-drop +3
tiktok-bold-yellowminimal-whitesubtle-dropkaraoke-gradientbig-serifserif-editorial
--watermark-scale
int默认20Watermark width as a percentage of the video width.
--watermark-url
asset_url默认Override the default Pipe2 logo with your own image (URL or local path). Empty means use the bundled Pipe2 watermark.
--watermark-variant
enum默认lightWhich bundled Pipe2 logo to use when --watermark-url is empty. "light" for dark videos, "dark" for bright ones; coloured variants match the clip palette.
amber · aqua · crimson +4
amberaquacrimsondarkemeraldindigolight
工作原理 6 个流水线
  1. transcription

    完整视频只转录一次，所有选中的短片都会复用同一份转录稿。

    输入来源: OpenClaw Creator, Why 80% of Apps Will Disappear
    文本 
    00:00:00,180  Today, I'm sitting down with Peter Steinberger, the creator of OpenClaw, the open source personal AI agent that has completely taken over the internet.
00:00:09,060  The GitHub repo exploded to over 160,000 stars practically overnight.
00:00:14,130  The community has built countless projects, like Malt Book, where bots talk among themselves.
00:00:19,560  And now, the bots are even renting humans to do tasks in the real world.
00:00:24,220  In our conversation, we discuss his aha moment, his contrarian development philosophies, and what this means for builders in 2026.
00:00:32,740  Let's dive in.
00:00:38,980  So good to see you, man.
00:00:39,960  Hey, what's up?
00:00:40,760  Um, so you've made something people want.
…
  2. highlights

    读取转录稿并挑选N个适合剪辑的时刻。若--clips提供了手动JSON列表，则跳过自动选片。

    JSON 
    [{"context":"Peter explains OpenClaw's core differentiator: local execution gives it access to everything the user can do, unlike cloud-based AI.","desired_seconds":32,"start_sec":95.24000000000001,"end_sec":127.64},{"context":"The vivid moment Peter realized OpenClaw's creative problem-solving: it autonomously transcribed a voice message using ffmpeg and OpenAI's API without being explicitly programmed to do so.","desired_seconds":106,"start_sec":515.182,"end_sec":621.252},{"context":"Peter's contrarian prediction: 80% of apps disappear because personal AI agents manage data and tasks more naturally than purpose-built applications.","desired_seconds":55,"start_sec":641.792,"end_sec":696.312},{"context":"Peter's philosophy on building: minimize friction by using Unix tools and CLIs instead of inventing new abstractions, letting the model handle creative problem-solving.","desired_seconds":54,"start_sec":1178.264,"end_sec":1232.584},{"context":"Peter articulates why swarm intelligence mirrors human society: individuals alone can't build iPhones or go to space, but groups specializing together achieve anything.","desired_seconds":46,"start_sec":258,"end_sec":303.5}]
  3. video-trim

    按句子边界裁出每个选定区间，并为字幕步骤返回与各短片准确对应的转录稿。

    视频
  4. video-reframe 仅当 --reformat

    设置--reformat后，会把每条短片转换为指定画幅，让当前说话者保持在画面内，并在镜头切换处自然调整构图；未设置时则保留原视频画幅。

    视频
  5. captions

    按照所选字幕样式和位置，把对应的转录内容添加到每条短片中。

    视频
  6. watermark

    在左上角叠加Pipe2标志（或你通过--watermark-url提供的标志），让内容被转发时仍保留出处；传入--no-watermark即可跳过。

    视频