Roadmap

As AI model accuracy improves and device performance advances, it's becoming feasible to run some models locally to solve problems previously considered challenging.

Language barriers will remain a major challenge for cross-regional communication in the foreseeable future.

DuRT will continue to be updated and maintained.

Task List

• UI optimization (Medium-high priority)
• Improve file transcription results (recognition accuracy, timestamps) (Medium-high priority)
• Enhance subtitle editing features (Medium priority)
• File-based Q&A and summarization AI capabilities (Medium priority)
• Add prompt management functionality (Medium priority)
• Support translation features on macOS (Medium priority)
• Use large language models to optimize overall process results (Medium priority)
• Support AppleScript automation (Medium-low priority)
• Add internal audio source selection (e.g., only recognize sound from specific apps) (Low priority) (Uncertain feasibility)
• Support non-streaming Whisper models with more language recognition (Extremely high priority)
• Add complete subtitle file generation for local videos/audio (High priority)
• Implement near real-time streaming speech recognition with Whisper (Medium-high priority)
• Add more translation methods (platform translation APIs, LLM APIs) (Medium-high priority)
• Support Apple Speech Recognition on macOS (Medium priority)
• Add service model management interface with decoupled services (Medium priority)
• Add microphone source selection (e.g., headset mic) (Medium priority)
• Implement automatic language detection using Whisper (no manual selection required) (Medium-low priority)

Discussion and Suggestions

If you have better ideas or suggestions, please Contact Us to communicate with us.

Open Source Acknowledgments

This project draws inspiration from many open source projects. Special thanks to:

• whisper.cpp
• whisper
• sherpa-onnx