| 开发者 | mwangip |
|---|---|
| 更新时间 | 2026年10月15日 00:45 |
| 捐献地址: | 去捐款 |
| PHP版本: | 8.0 及以上 |
| WordPress版本: | 7.1 |
| 版权: | GPLv2 or later |
| 版权网址: | 版权信息 |
[mwcap_video id="123"] shortcode and a Gutenberg block/audio/transcriptions endpoint. No provider is preselected: you enter the base URL and API key of the service you choose (for example OpenAI, Groq, or your own Whisper server). When you press Generate, the audio of the selected video (or the video file itself when FFmpeg is unavailable), the chosen language and your API key are sent to that service. See the provider's terms and privacy policy. You must confirm this each time.proc_open() using an argument array (no shell), and only with the binary path an administrator configured. If your host disables proc_open(), the plugin falls back to sending small MP4/WebM files directly to the provider.
Background jobs use WP-Cron. If DISABLE_WP_CRON is set, trigger wp-cron.php from a system cron. See System Status.
MP4 or WebM files below your provider's size limit (typically 25 MB for hosted providers). MOV and larger files need FFmpeg.
Captions need accurate segment and word timestamps from a speech-to-text (transcription) model. The AI Client bundled with WordPress 7.x covers text, image, speech-generation, music, video and embedding tasks, but has no transcription capability, so it cannot return timestamped transcripts. The plugin keeps a provider interface (mwcap_transcription_provider filter) so a core-based provider can be added once transcription is available.