Voice input

Control speech-to-text dictation for your organization

Written By Lina Pridat Hauschild

Last updated 1 day ago

Overview

Voice input lets team members speak a message instead of typing it. Noxtua records the clip in the browser, transcribes it, and places the text in the chat box for review before anything is sent. The same microphone also powers Dictate a skill in the skill editor.

How voice input is turned on

Noxtua enables speech-to-text for your organization.
Your admins then see a Voice input switch under Settings → Admin → Workspace and can turn it on or off.
When it is on, everyone in the organization sees the microphone button and can dictate.
When it is off, the microphone button is hidden for everyone.
If the switch is greyed out, Noxtua has not enabled speech-to-text for your organization yet. Contact Noxtua support.

Why it is governed centrally. Spoken input is sensitive: a lawyer dictating a client matter is speaking privileged information aloud. Voice input is therefore designed so that:

No audio is stored. The recording is sent for transcription and then discarded; only the resulting text is kept as part of the chat.
Nothing is sent automatically. The transcript appears in the chat box, where the user can edit or delete it before sending.
Transcription runs on Noxtua's own speech-to-text service, never a fallback vendor. If the service is not configured, voice input stays off rather than routing audio elsewhere.

Who can change it

Noxtua enables speech-to-text for your organization. Admins can then switch voice input on or off for the whole organization under Settings → Admin → Workspace. Individual users cannot change it themselves.

Role	What they can do
Noxtua (operations)	Enables speech-to-text for your organization, which makes the Voice input switch available to admins
Organization admin	Switches voice input on or off under Settings → Admin → Workspace
IT / browser administrator	Can block microphone access in managed browsers, so recording fails on those devices
Team member (with app access)	Uses voice input once it is on; grants or denies the browser's microphone prompt

Options explained

On — The microphone button appears in the chat box, and Dictate a skill appears in the skill editor for everyone with app access. Recordings are transcribed by Noxtua's speech-to-text service, and the audio is discarded.
Off — The microphone button and Dictate a skill are hidden for everyone. No audio can be recorded or sent. Typing is unaffected.
On, microphone blocked by browser policy — Voice input is on for the organization, but managed browsers deny microphone access. Affected users see the button, and recording fails with Voice input failed.
Not available (switch greyed out) — Noxtua has not enabled speech-to-text for your organization yet, so admins cannot turn voice input on. Contact Noxtua support.

How to set it up

Once Noxtua has enabled speech-to-text for your organization, admins turn voice input on or off under Settings → Admin → Workspace → Voice input. After changing it, confirm the change in the chat box.

Switch it on or off

Open Settings and, under Admin, select Workspace.
Turn the Voice input switch on or off. The change applies to everyone in your organization.
If the switch is greyed out, or saving shows Voice input isn't available for your organization, Noxtua has not enabled speech-to-text for your organization yet. Contact Noxtua support.

Confirm it took effect

Reload Noxtua in your browser. The change can take a few minutes to reach everyone; if nothing has changed, reload again after 5 minutes.
Open a new chat. When voice input is on, a microphone button labeled Start voice input appears in the chat box.
Click it and allow microphone access when the browser asks. The bar reads Listening …. Click Done, and the bar shows Transcribing … before your text appears in the chat box.

If the button does not appear after reloading, either an admin has not turned voice input on, Noxtua has not enabled speech-to-text for your organization yet, or your browser cannot record audio. If clicking the button shows Voice input is turned off for your workspace., the change has not reached your session yet: wait a few minutes and reload. If it shows Voice input is not configured., contact Noxtua support.

Effect on users

Users are not notified when voice input is turned on or off; the microphone button appears or disappears the next time they load Noxtua, and existing chats are unchanged.

Dictate a skill rewrites the recording into a structured skill description instead of pasting it verbatim. If the field already has text, the user chooses Replace or Append.
Supported browsers: current Chrome, Edge, Firefox, and Safari. Browsers that cannot record audio do not display the button.

Message the user sees	What it means
Voice input failed.	The microphone could not open (permission denied or blocked by policy), or the network or transcription service failed
That recording could not be transcribed. Try recording again.	The clip was empty, unreadable, or too large; recording again usually fixes it
Voice input is not configured.	Noxtua has not enabled speech-to-text for your organization yet; contact Noxtua support

Warning

Recordings cannot be recovered. Because audio is never stored, a recording that is discarded, canceled during transcription, or fails to transcribe is gone for good, and neither you nor Noxtua support can retrieve it. Ask team members to keep dictations short and to review the transcript before relying on it.
Sent transcripts are standard chat content. Turning off voice input later does not remove text that was already dictated and sent. To remove it, delete the chats in question under your retention settings.

FAQ

Can I enable voice input for some users or teams only? Not currently. Voice input is either on or off for the entire organization. To restrict it to specific devices, block microphone access through your browser management policy.

Is audio recorded or stored anywhere? No. The clip is sent only once to Noxtua's speech-to-text service and discarded after transcription. Only the transcript is kept, and only if the user sends it as part of a message.

Is the transcript sent to the AI model right away? No. It is placed in the chat box, and the user decides whether to edit, send, or delete it.

Which model is used when a dictated message is sent? The same model as for any typed message: the one selected in the chat box's Model selector. Voice input does not change which model receives the text.

Which languages are supported? Language support depends on Noxtua's speech-to-text service. Confirm with your account manager which languages your team relies on before rollout.

Is there a length limit? Recordings are capped at 25 MB per clip, which is enough for many minutes of speech, and transcription can take up to 2 minutes. Long dictations are best split into several shorter ones.

A user says the microphone button is missing. What should I check? First, check that voice input is switched on under Settings → Admin → Workspace. If the switch is greyed out, contact Noxtua support. Then ask the user to reload the page and use a current browser. If the button appears but recording fails, check the browser's site permissions and any managed microphone policy.

Was this helpful?