Voice
Dictate messages, hear answers read aloud, talk hands-free, choose the reading voice, and teach Nexa how you write.
Voice covers two separate things in Nexa Agora. Speaking and listening: you dictate, answers and documents are read aloud, and you can hold a hands-free conversation. Your writing voice: on the My Voice shelf, you teach Nexa how you write, so agents drafting for you sound like you. This page shows you both.
Before you start
- Speaking and listening are part of the Pro and Enterprise plans. On other plans, the voice controls stay visible but greyed, with a small Pro.
- Writing voices on the My Voice shelf are for everyone, and private to their owner.
Dictate a message
In Silvestro, or in an Agora discussion, click the voice button in the chat box.
Choose Dictate a message and speak.
Click the voice button again, Stop and transcribe.
The page shows Transcribing…, then the text is in the chat box.
Edit it and send.
On My Desk, the Dictate button in the toolbar does the same in your notebook. Click Stop & insert to place the text.
Hear the answers
- To hear every finished answer, open the voice button and switch Spoken replies on.
- To hear one message, click Listen on it. Use Pause and Resume. Messages in Agora have Listen too.
- To hear a document, click Listen in the document's panel. A line shows the reading voice, with Change it.
Talk hands-free
Open the voice button and choose Hands-free conversation.
You speak, pause, and hear the answer, with no clicks between turns.
Follow the chip that shows the state: Listening — pause to send, Thinking… or Speaking — talk to interrupt.
To stop, open the voice button and choose End the voice conversation.
Choose the reading voice
Quickly: open the voice button, then Reading voice. Pick Auto — follow the language of the text, or a voice. Click Hear … speak to preview it.
In full:
Open Governance › Settings › Voice.
Under Reading mode, choose:
With Auto, choose a Fallback voice for a language without a native voice: Nexa's default voice, or one you pick.
Browse the Voice catalog. Search with Search voices or languages…, filter by Language, Accent, Gender, Quality, Latency and Multilingual, and use the tabs Recommended, Favorites, Recent, Best multilingual and All voices.
Click a voice to see its portrait, Preview, its languages, quality and latency. Click Use this voice, or Add to favorites.
Teach Nexa how you write
The My Voice shelf keeps the ways you write as sealed profiles. Any agent drafting for you writes to the seal, and tells you exactly what it checks.
Open Focus › My Desk, then My Voice shelf.
Click New voice, or Teach it your voice.
The page offers three ways in: Three doors. One room.
Answer the questions about Presence, Expression, Format and The seal.
YOUR SPECIMEN rewrites itself as you answer. Traits observed in a pasted sample are already filled in: tap to confirm, or choose otherwise.
Read the two test pieces written in your voice. Confirm them, or click Something is off — adjust.
Seal the voice.
It appears on the shelf with THE CONTRACT: what is CHECKED ON EVERY DRAFT — WITHOUT EXCEPTION, such as your banned words, emoji, hashtags and bullets, who speaks and whether there is an ask; and what GUIDES THE PEN.
On a voice card, you can Edit, Rename or Remove voice. Drafts already written keep their text.
Write with a sealed voice
In Silvestro's chat box, or in the Document Editor, click the Voice chip, Write in your voice.
Pick a voice.
The chip shows the voice's name, for example LinkedIn — Sara; its tooltip reads Writing in: and the name. Every draft in that conversation is written to its contract.
Manage voices on My Voice shelf → takes you to the shelf.
What you will see
Good to know
- The fallback voice is saved on your account. The chosen voice, favorites and recent voices are saved on the device you are using.
- Every voice costs the same.
- If your browser cannot record, Dictate and Hands-free are not shown, but spoken replies and the reading voice are.
- One person can keep several writing voices, for example for LinkedIn, a newsletter and formal documents, each sealed with its own mark.
- The documentation in Settings › About can also be read aloud with Listen, when voice is available.