Voice Input

CRHQ supports voice input: record your message with your microphone, and it's automatically transcribed to text and sent to the agent.

How to Use Voice

1. Click the microphone icon in the message input area
2. Speak your message
3. Click Stop when done
4. The recording is transcribed automatically
5. Review the transcription and click Send

Transcription

Audio is transcribed using Groq's Whisper model, fast and accurate. The transcription appears in the message input so you can review and edit before sending.

Audio Device Selection

If you have multiple microphones, click the dropdown arrow next to the microphone icon to select your preferred input device. CRHQ remembers your selection.

Speech to Text Through the API

Products that use a limited API key can turn audio into text with POST /api/transcribe: up to 30 minutes of audio a day per server on CRHQ's shared key, and no daily limit once you add your own Groq key in Settings > Credentials > Groq (field api_key). The microphone in the app is not affected.

Any Groq key saved under Settings > Credentials > Groq counts as the server's own, so API transcription then uses it with no daily limit. Without one, the 30-minute daily allowance is shared by all API keys on the server.

Tips

• Speak clearly and at a natural pace
• You can edit the transcription before sending if needed
• If transcription fails, you can retry without re-recording
• Voice input works on both desktop and mobile browsers
• Browser permission for microphone access is required on first use