# Voice input

You can use your microphone to control the agent. Simply click the microphone icon, close to the send button, and start dictating.

## Browser considerations

Note the voice input is fully controlled by your browser, which implies:

* The browser uses your default microphone. You can change which microphone to use either in your browser settings or in your Operating System.

* The browser uses your default language. You can configure your default language in your browser settings.

* The browser may use remote services for speech recognition. Some browsers (typically Chromium based browsers like Chrome, Edge, etc.) offer the ability to perform on-device recognition for certain languages. You can choose this option in Tidewave settings.

* Not all browsers have speech recognition available (neither remotely or on-device).

## Tips

Here are some useful tips for making the best of voice input:

* While using the voice input, you can still use the [Inspector](inspector.md) or click on lines during [Code Review](code_review.md), and Tidewave should correctly inject them as part of your speech. However, some browsers require you to briefly pause before clicking.

* Use `Ctrl+Shift+M` (or `Cmd+Shift+M` on macOS) to quickly toggle voice input.