PasteFerry (隔空贴): clipboard and file sync across macOS, Windows and Linux. Direct-first, end-to-end encrypted. Optional official access: no server to host. Independent self-hosting remains available.
PasteFerry 2.1 portrait launch film
A 25.5-second, 1080×1920, 30 fps video with Chinese narration, burned-in captions, a matching SRT, a portrait cover and original synthesized music. No external media or model service is used. The computer screens are labeled motion demos, not recordings of a user's clipboard. All information uses the public website; no administrative hostname, test identity or credential is included.
The opening immediately shows a cross-network copy/paste result. The next cuts show the official-service switch, direct-first routing, the current relay allowance and the official download address. Text stays above the bottom platform overlay area and clear of the right action rail. Do not claim unlimited relay or that private production storage is proven by the public source.
Requirements: macOS with the Tingting voice (say), Chrome, Node.js,
puppeteer-core, Python 3 with NumPy, and FFmpeg with libx264. Set NODE_PATH to
the puppeteer installation, and optionally NODE_BIN, PYTHON_BIN, CHROME.
./tools/shorts/build.sh /absolute/output/path
Each film lives in its own directory with story.json and video.html; pass it
as the second argument (the 2.1 film in this directory is the default). The 2.10
themes film is in themes-2.10/; its prepare.sh renders the app screenshots it
needs with ui/tool/screenshots, so it also needs Flutter and fonttools.
./tools/shorts/build.sh dist/shorts-2.10 tools/shorts/themes-2.10
The narration voice is set by tts in story.json. Without it, the local macOS
Tingting voice is used. The themes film uses {"engine": "edge"}: a Microsoft
neural voice (zh-CN-YunxiNeural, the young male voice common on Douyin;
zh-CN-YunjianNeural is a calm mature male voice) through the
edge-tts Python package, which sends the narration text to Microsoft's online
service. Override the voice for one render with TTS_VOICE. Check the voice's
terms before commercial use.
Use node tools/shorts/render.js /output/path stills to inspect the key frames
before rendering. story.json owns the narration, timings and captions. Visuals
in video.html use the same deterministic timeline. Refresh both when editing.
Build outputs stay outside the 15 release attachments, and outside dist/,
which scripts/build.sh clears.