Real-Time Speech Recognition & Transcription for macOS (R&D Proof of Concept) [In Development]

This project was developed as a proof of concept to explore the capabilities of real-time speech recognition and natural language processing on macOS.

Utilizing Apple's Speech framework, it captures and transcribes spoken input in real-time. Additionally, the integration of the Natural Language framework allows for tagging and semantic analysis of the transcriptions, enabling more nuanced and meaningful data extraction.

The goal of this R&D effort was to investigate the potential for building intuitive and responsive voice-driven applications that leverage built-in macOS technologies. The findings from this project can serve as a foundation for more complex applications, such as automated transcription services, or accessibility tools.

Further development could explore enhancements like improved contextual understanding in tagging.

Features

• Real-Time Speech Recognition:
  • Capture and transcribe speech in real-time
• Natural Language Tagging:
  • Automatically generate tags for saved transcriptions
• Transcription Management:
  • Save, edit, and delete transcriptions from a split view interface

Prerequisites & Installation

• macOS Sequoia 15.0 or later
• Xcode 16+
• Clone, open and build the project in Xcode

Usage

Recording Audio

1. Click the microphone button to start/stop recording; the visualizer will display the audio levels in real-time.
2. Click the Save button below the transcribed text; this will save the transcription to the list of recordings.
3. After saving, the app will switch to the split view, where you can edit your transcription and tags.

Managing Transcriptions

1. On the left side of the split view, you'll see a list of your saved transcriptions.
2. Click on any transcription to view and edit its details on the right side of the split view.
3. You can edit the transcription title, content, and tags. Click the Save button to update the transcription.
4. To delete a transcription, click on the trash icon in the toolbar.

Warning

When there is no audio input device connected, the app may not function as expected (you can obviously use your built-in microphone on your macOS machine but if you're using a Mac mini, for example, I highly recommend connecting using your AirPods or some other input source).