zshot/cliDownload

--navigator-video-shows-captions

CLI
--navigator-video-shows-captions
LicensePro

Draws explanatory captions onto the --navigator-captures-video recording. When the navigator judges that the reason for an action helps explain how it is reaching the goal, it writes a caption, drawn under the action — named as the navigator calls it or else as the page does, such as Click “Sign in” — and held before the action runs. The video never shows a dialog, so the navigator captions each dialog it answers, quoting it. The caption is in the language of the goal, as is the action line in English, German, Spanish, French, Italian, Portuguese, Japanese, Korean, Chinese, and Arabic; other languages get an English action line. Filled values stay out of both. Routine steps carry no caption and no hold.

Each caption is held for one second to notice it plus the time an adult takes to read it silently, from 2 to 15 seconds: English at 238 words a minute, with longer words and numbers taking longer, and Japanese, Chinese, Korean, and Arabic at their own measured rates. A caption too long for its place scrolls through its hold.

Captions are drawn with the Noto fonts that ship with ZShot, so they read the same on every host, in Latin, Cyrillic, Greek, Arabic, Japanese, Korean, and Chinese scripts. --navigator-video-captions-position places them. A typesafe navigator writes no captions.

Example

Run locally
zshot -f pipeline.png \
  --navigator-type "anthropic://YOUR_API_KEY@" \
  --navigate "Open the pipeline" \
  --navigator-captures-video session.webm \
  --navigator-video-shows-captions \
  https://zshot-cli.com/example_assets/zrm/