Local Transcription

Podex transcribes downloaded podcast audio with Apple's Speech framework. The v0 goal is local English transcription that can complete without sending audio to a server.

Before You Transcribe

• Download the episode audio first.
• Confirm the device has the English speech asset installed.
• Keep enough free storage for intermediate data and transcript output.
• Use a physical device for the most representative Speech behavior.

Simulator builds are useful for UI development, but local Speech asset behavior and long audio transcription should be verified on supported hardware.

During Transcription

Podex tracks transcription progress and saves timestamped transcript segments. The transcript view can follow playback position, highlight the current segment, and jump playback when you select a segment.

If transcription is interrupted, Podex should preserve durable state and let you retry without presenting partial output as complete.

When Local Speech Is Unavailable

On-device recognition depends on Apple's current support for the device, OS, locale, and installed model assets. When local recognition is unavailable, Podex reports that condition instead of automatically switching to server recognition.

For troubleshooting, see Support.