whisper

name	whisper
description	Audio transcription (Whisper) and text-to-speech (OpenAI TTS). Use for processing audio messages or generating voice responses.

Whisper

Listen to audio, say things out loud. Powered by OpenAI's Whisper and TTS.

Setup

Requires OPENAI_API_KEY in your .env file.

Quick Reference

# Listen to audio (transcribe)
clappie whisper listen /path/to/audio.ogg

# Say something (text-to-speech)
clappie whisper say "Hello world"

# Say with specific output path
clappie whisper say "Hello world" -o recall/files/hello.opus

# Pick voice by mood
clappie whisper say "I'm feeling chaotic" -m chaotic

# Use specific voice
clappie whisper say "Serious business" -v onyx

# List all voices
clappie whisper voices

Storage

Audio files go to recall/files/ with everything else. Override with -o if needed.

Voices

Voice	Vibe
alloy	Neutral, balanced
echo	Male, warm
fable	British male, expressive
onyx	Deep male, authoritative
nova	Female, friendly
shimmer	Soft female, gentle

Mood-Based Selection

When you use -m mood, a matching voice is picked:

• excited: nova, fable, alloy
• chill: shimmer, echo, alloy
• serious: onyx, echo
• playful: fable, nova, alloy
• chaotic: fable, nova
• existential: shimmer, onyx, echo

Integration with Telegram

When receiving voice messages in a sidekick:

# Listen to what they said
clappie whisper listen recall/files/voice-xxx.ogg

# Say something back
clappie whisper say "Your response here" -m playful
# Output lands in recall/files/say-<timestamp>.opus
clappie sidekick send-file voice recall/files/say-<timestamp>.opus

Notes

• Whisper supports: mp3, mp4, mpeg, mpga, m4a, wav, webm, ogg
• TTS outputs opus format (great for Telegram)
• Speed can be adjusted with -s (0.25 to 4.0)