Frequently Asked Questions (FAQ)

Find quick solutions and troubleshooting tips for common issues in DuRT.

1. Permissions & Audio

Q: Why is there no sound captured during Live Caption?

Ensure that Screen & System Audio Recording is enabled for DuRT in System Settings ➔ Privacy & Security.
Make sure the audio playing app (e.g. Safari, Chrome, Zoom) is not muted.
If you recently granted permissions, please restart DuRT (⌘ + Q).

Q: Does DuRT require kernel extensions or virtual audio drivers?

No. DuRT uses Apple's native ScreenCaptureKit and CoreAudio, requiring zero kernel extensions or third-party virtual audio cables.

2. Models & Performance

Q: Does Apple Speech Recognition work completely offline?

Yes. When on-device dictation is enabled in macOS System Settings, Apple Speech functions 100% offline on your device with zero internet access required.

Q: How can I achieve higher accuracy or specialized multilingual transcription?

We recommend integrating the OneASR Gateway or OpenAI Whisper API. These engines support advanced word-level timestamp alignment and speaker diarization across dozens of languages.

3. Subtitles & Workflows

Q: How do I export subtitles with word-level timestamps?

In the Media Transcription step configuration, enable the Word-level Timestamps toggle. When export is executed, select .json or timed .srt.

Q: Can I run custom prompts for LLM proofreading?

Yes. In the LLM Process step, you can customize the system prompt or select from built-in categories (Proofreading, Translation, Summary, Semantic Segmentation).

4. Account & Billing

Q: Is DuRT free?

The core application and macOS built-in Apple Speech Recognition are free to use.
When connecting to cloud providers (OpenAI, DeepSeek, Alibaba Cloud), you use your own API keys and pay only the provider's standard rates.