Grok Caption

A small Windows-friendly live caption page backed by xAI Grok Speech-to-Text.

Run

$env:XAI_API_KEY="xai-your-key"
node .\server.mjs

Open http://localhost:8787.

Notes

• Use 屏幕/标签页音频 for video calls, streams, or browser audio. In the browser share picker, choose a tab/window/screen and enable audio sharing when offered.
• Use 麦克风 for spoken input near the computer.
• This version sends short 16 kHz WAV segments to https://api.x.ai/v1/stt, then optionally translates them through https://api.x.ai/v1/chat/completions.
• The default segment length is 3 seconds; use 2 seconds for lower latency or 5-7 seconds for fewer requests.
• xAI's documented formatting languages do not currently list Chinese. Leave language on 自动 for unsupported languages.
• Translation defaults to Simplified Chinese. Set 翻译到 to 不翻译 for raw transcription only.