Voice Input

What Voice Input Does

Aori can transcribe your voice messages and treat them as text input, allowing you to interact with your agent hands-free.

• In-app Chat: Hold the microphone button in the Aori chat screen.
• Messaging Channels: Send a voice message through Telegram or WhatsApp, and Aori will transcribe it automatically.

How it Works

Aori sends the audio to a speech-to-text (STT) backend for transcription. The resulting text appears in the input field, allowing you to review and edit it before Aori responds.

STT Backends

To use voice input, you must configure a speech-to-text provider in Settings → Connections → Voice → STT Backend.

• Groq Whisper (Default): Fast, high quality, and free. Uses your existing Groq API key.
• OpenAI Whisper: An alternative using your OpenAI API key.

In-App Push-to-Talk

1. Open the Chat screen in the Aori app.
2. Hold the microphone button next to the text input.
3. Speak your request and release the button to send for transcription.
4. Review the transcribed text and tap Send to communicate with Aori.

Voice via Telegram/WhatsApp

If you have connected a channel, simply send a voice message to your bot or linked session. Aori will transcribe it and respond as if you had typed the message.

Note

Text-to-Speech (Voice Output)

Aori v1.3 now supports voice output (Aori speaking back to you) via Edge TTS (free) and ElevenLabs (premium). You can enable this in Settings → Connections → Voice.

Troubleshooting

• Microphone Not Showing: Ensure that Aori has permission to access your device’s microphone in your OS settings.
• Poor Quality: If transcription is inaccurate, try switching your STT backend in Settings → Connections → Voice.