Voice Input

What is Voice Input?

Voice Input lets you talk instead of type. Wherever you see the microphone button in a composer, you can click it, speak, and ScoutSocial transcribes what you said and inserts it into the text field. Transcription runs through OpenAI's Whisper model, which automatically detects the language you speak and returns the text in that same language, so there is no language picker to set. It is a quick way to draft a prompt or add text hands-free.

Where to find it

Voice Input is not a separate page. It is a microphone button that appears inside several composers:

• Create Posts composer - the mic sits in the prompt bar when the field is empty (the main Generate / Create Posts screen).
• Refine with AI - the mic appears in the refine chat drawer where you describe the changes you want.
• Brainstorm composer - the mic is in the composer on the Whiteboard brainstorm surface.
• Video editor Text tool - the mic is available when adding or editing text in the video editor.

The microphone shows when the field is empty and not busy. Once you start typing, the field's own submit control takes the button's place.

How it works

1. Click the microphone button in the composer.
2. Your browser asks for microphone permission the first time. Allow it to continue.
3. The button turns into a red, pulsing stop control while it records. Speak your text.
4. Click the button again to stop. It shows a spinner while Whisper transcribes the clip.
5. The transcribed text is appended to whatever is already in the field, separated by a single space.

A few behaviors worth knowing:

• Auto language detection - Whisper detects the spoken language from the audio and transcribes in that language, so you can dictate in the language you actually speak without changing any setting.
• Recording length - a single dictation is capped at 60 seconds. If you leave it running, it stops on its own and transcribes what it captured.
• Nothing is stored - the audio is sent for transcription and is never saved. Only the resulting text lands in your composer.
• Append, not replace - dictation adds to the current field contents rather than overwriting them, so you can mix typing and speaking.

Availability

Voice Input relies on two things:

• A supported browser. The button works anywhere microphone recording is available (recent Chrome, Edge, Firefox, and Safari 14.1 and later). On unsupported browsers the mic is disabled with a tooltip that says voice input is not supported. Voice input also needs a secure (https) connection; over plain http the browser blocks the microphone.
• A configured OpenAI provider. Transcription uses Whisper, which is only available through OpenAI. If your workspace has no OpenAI provider set up (for example, an Anthropic-only setup), the button is present but transcription returns a message asking you to configure OpenAI in Settings > AI Providers.

If the browser blocks the microphone, a message points you to the mic/camera icon in the address bar (or Site settings) to allow the microphone for the site, then try again.

Tips

• Speak in short, clear phrases and pause the recording between thoughts. You can dictate multiple times and each transcript appends to the field.
• If nothing appears after you stop, check that the microphone is allowed for the site in your browser's address bar.
• Use Voice Input to rough out a prompt quickly, then edit the transcribed text by hand before generating.

Related

• Create Posts (AI Generate) - the main composer where the mic appears.
• Collaboration and Brainstorm - the Whiteboard brainstorm composer that also supports dictation.