Frequently Asked Questions (FAQ)

A comprehensive guide to common questions, technical troubleshooting, and configurations for DuRT 3.0.

---

1. Permissions & Audio Capture Troubleshooting

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

• Verify System Recording Permissions: Go to macOS System Settings ➔ Privacy & Security ➔ Screen & System Audio Recording and ensure DuRT is toggled on;
• Check Audio Source Volume: Confirm that the playback application (e.g. Safari, Chrome, Spotify) is not muted;
• Restart DuRT: If permissions were just granted, fully quit DuRT (⌘ + Q) and relaunch it.

Q: Does DuRT require virtual audio drivers or kernel extensions? Will it conflict with conferencing apps?

• Zero virtual audio drivers required. DuRT is built natively on Apple’s modern ScreenCaptureKit and CoreAudio frameworks. It captures system sound safely inside the macOS sandbox without injecting kernel extensions (no BlackHole or Soundflower required), eliminating audio device contention with Zoom, Teams, or AirPods.

Q: How can I prevent speaker echo when using the microphone?

• DuRT features built-in hardware/DSP Acoustic Echo Cancellation (AEC). If you are recording dictation while playing audio over Mac speakers, enable “Echo Cancellation” in audio settings or wear headphones for pristine vocal isolation.

---

2. Engines, Models & On-Device Performance

Q: Can Apple Speech Recognition work completely offline without internet?

• 100% offline. When local speech models are downloaded in macOS System Settings ➔ Keyboard ➔ Dictation, Apple Speech operates entirely on your Mac's Neural Engine. Zero audio or text data leaves your device.

Q: How can I achieve higher accuracy, specialized multilingual recognition, or custom vocabularies?

• For complex multi-speaker meetings or specialized terminology, connect the OneASR Gateway or configure OpenAI Whisper API. Cloud and gateway models provide industry-leading word-level timestamp precision and unsupervised speaker diarization.

Q: What is the typical latency for real-time bilingual live captions?

• Thanks to our streamlined streaming pipeline, live caption latency with macOS 15 on-device neural translation or DeepSeek / OpenAI APIs is typically between 200–500 ms, ensuring seamless comprehension during international conferences and video streaming.

---

3. Visual DAG Workflows & Studio Video Export

Q: What is a DAG workflow? How does it differ from traditional serial pipelines?

• DuRT 3.0 departs from rigid step-by-step linear chains in favor of a Directed Acyclic Graph (DAG) 2D canvas architecture.
• By wiring nodes via bezier curves, you can execute parallel branch fan-outs: for example, a single transcription output can simultaneously feed subtitle formatting, multilingual TTS dubbing, and meeting minutes synthesis in parallel, dramatically cutting total processing time.

Q: Why can't I connect certain steps on the canvas?

• DuRT enforces a Strongly-Typed Data Contract (DuWorkflowDataType). A node outputting audioStream can only connect to steps accepting audio input (such as ASR). The built-in rule engine validates connections to prevent runtime type errors.

Q: What styling controls are available for hardcoded burn-in subtitles?

• In the Studio editor, you can customize font families, font sizes, weights, multi-line wrapping, outline stroke width, text drop shadows, and semi-transparent background padding. Videos can be rendered and exported at up to 4K UHD resolution with hardware acceleration.

---

4. Cloud Storage & Automated Multi-Platform Delivery

Q: Which cloud storage and knowledge platforms does DuRT support?

• DuRT 3.0 natively connects to 6 major storage and integration platforms:
  1. S3-Compatible Object Storage: AWS S3, Cloudflare R2, MinIO, Alibaba Cloud OSS, Tencent COS (supports credential-free presigned uploads);
  2. WebDAV Cloud Drives: Nutstore, Nextcloud, Synology WebDAV Server;
  3. NAS / Local Folders: Direct local filesystem paths or mounted SMB/NFS network shares;
  4. Notion: Automatic structured database page creation with native Notion blocks;
  5. Obsidian: Direct write to local Vaults with standardized YAML front-matter;
  6. Custom HTTP Webhook: JSON payloads delivered to custom REST APIs or Zapier / n8n pipelines.

Q: Does Notion integration require complex developer app reviews?

• Not at all. DuRT leverages Notion's official Internal Integration framework. You simply generate an internal secret token in Notion's developer portal and connect your target database in under one minute.

---

5. Account, Billing & Early Supporter Perks

Q: Is DuRT Free really free? Are there duration limits?

• Core dictation and transcription are permanently free: DuRT Free features unlimited audio duration, local media file transcription, on-device Apple speech recognition, word timestamps, and speaker diarization without artificial 10-minute limits.
• DuRT Pro: Unlocks the visual DAG canvas, AI semantic refinement and translation, TTS character dubbing, video burn-in subtitles, 4K rendering, and 6 automated cloud delivery channels.

Q: Are there extra fees when using my own API keys (BYOK)?

• Using Apple on-device recognition or local Ollama models is 100% free with zero network costs.
• When connecting to cloud providers (OpenAI, DeepSeek, etc.), requests go directly from your Mac to the provider. Token usage is billed directly by the provider according to your account with zero middleman markup by DuRT.

Q: What perks are available for early adopters who purchased prior to the price adjustment?

• Founding Lifetime Buyers: Choose between claiming 1 free Mac App Store Lifetime Promo Code for a friend, or registering to receive a free Lifetime License for our upcoming mobile or ecosystem companion app upon launch;
• Subscribers Exceeding 15 Months: In addition to permanent Mac Pro access, subscribers with over 15 months of support will also receive gifted membership duration / perks when our mobile companion launches!
• See the full Legacy User Benefits Guide.