goodroot/hyprwhspr v1.30.0

on GitHub

Highlights

• Added OpenAI gpt-realtime-whisper support with live mic OSD transcript preview.
• Reworked Realtime WebSocket setup into a simpler provider + model selection flow.
• Realtime WebSocket is no longer labeled experimental.
• Added automated Continue PR review through opub.
• Updated website dependency devalue from 5.6.4 to 5.8.1.

User-Facing Changes

• Realtime setup is shorter and less confusing.

• Selecting Realtime WS now shows a flat list like:

  OpenAI: GPT Realtime Whisper
  Google (Gemini): Gemini 3.1 Flash Live
  Google (Gemini): Gemini 2.5 Flash Native Audio
  ElevenLabs: Scribe v2 Realtime
  Custom WebSocket endpoint

• Setup no longer asks users to choose Transcribe vs Converse.

• Realtime WebSocket setup now defaults to transcription mode.

• OpenAI realtime transcription can show partial text in the mic OSD before final paste.

• Realtime WebSocket docs now describe OpenAI, Google, and ElevenLabs support as mature.

Developer / Maintenance Changes

• Added realtime preview tests.
• Added mic OSD daemon safety tests.
• Added startup guard tests around realtime partial callback registration.
• Added TRANSCRIPT_PREVIEW_FILE runtime path.
• Added pycairo to Python requirements and distro install docs/scripts.
• Updated config schema for realtime_transcription_delay.

Full Changelog: v1.29.4...v1.30.0