🎙️ Real-time Meeting Recognition

Real-time mode is ideal for ongoing meetings, lectures, or online video calls.

Last Updated: 2026-03-31 · Language: English

1. Audio Input Sources

Owl Meeting provides versatile audio capture schemes for different office scenarios:

• Microphone: Focuses on capturing local speakers, suitable for lectures, personal recordings, or noisy environments.
• System Sound: Directly captures internal computer audio, suitable for transcribing web videos, podcasts, or remote meetings.
• Dual Mode: Captures both microphone and system sound. This is recommended for remote meetings (like Microsoft Teams, Zoom, Slack) to record both sides.

2. Low-Latency Mode

For high-performance office needs, you can enable "Low Latency Mode":

• Streaming Feedback: The AI engine outputs dynamic drafts at high frequency, significantly reducing perceived latency.
• Real-time Correction: Text is continuously revised based on context, providing a "what you hear is what you get" experience.

3. Captions Mode & Voice Input

Transcription results can be integrated into your workflow in several ways:

• Captions Mode: A transparent floating window on your desktop. Perfect for reading real-time subtitles while watching live streams or video tutorials.
• Voice Input: An efficiency tool that types recognized text directly into any active text field (Word, Email, Chat, etc.) like an IME.

4. AI Tasks

Before transcription starts, you can preset automated AI processing chains:

• Real-time AI: AI tasks (translation, summary, polishing) are executed simultaneously with the transcription.

5. Troubleshooting Tips

• Missing segments? Lower the [Voice Detection Threshold] in settings.
• Captions too long? Lower the [Max Voice duration] in settings.
• Latency increasing? Check CPU usage or lower the [Max Voice duration].
• Traditional Chinese output? Set [Chinese Conversion] to [Traditional Chinese] in Settings.