Implemented audio utilities module: **Created files:** - `src/librechat_audio/utils/__init__.py` - Package exports - `src/librechat_audio/utils/audio.py` - Audio format utilities **Key functions:** - `normalize_format()` - Normalizes format strings (case, leading dots, whitespace) - `is_supported_input_format()` - Validates STT input formats (flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, webm) - `is_supported_output_format()` - Validates TTS output formats (mp3, opus, aac, flac, wav, pcm) - `get_content_type()` - Returns MIME type for audio format - `get_file_extension()` - Returns file extension with leading dot - `get_audio_format()` - Detects format from magic bytes or filename **Test results:** - 73 new tests in `tests/test_audio_utils.py` - all passing - Full test suite (191 tests) - all passing **Verification:** - All format constants match OpenAI Audio API spec in `docs/openai-audio-api-spec.md` - Magic byte detection tested for MP3, FLAC, WAV, OGG, WebM