Skip to content
Product · Voice

Say it once. Your workspace writes it down.

Talk into the text boxes you write in and get clean, ready-to-send text. Correct yourself mid-sentence, rewrite a paragraph by asking, hear your agents answer out loud, and turn a meeting into a summary and the tickets that come out of it.

3 s to take a message back7 one-phrase rewrites200 workspace words
Speak naturally

The edit is the feature, not the transcript.

Filler words go. Punctuation lands where you meant it. Change your mind halfway and the correction is what gets sent.

“um can you move the review to Thursday no wait Friday afternoon”

Can you move the review to Friday afternoon?

“mention Priya the invoice total is 1200 comma not 1100 new paragraph thanks”

@Priya the invoice total is 1200, not 1100.

Thanks

“the client wants the whole thing redone by Monday” … “make it more formal”

The client has asked for the full deliverable to be revised by Monday.

Dictation everywhere

Almost every text box has a microphone.

Chat, tickets, knowledge base pages and forms. Click the mic that appears in the field, or hold your push-to-talk key and speak.

  • A live level meter and the words heard so far, so you know you are coming through.
  • The microphone starts listening the moment you press, so your first word is not lost.
  • Pick which microphone to use. Each browser remembers it.
  • Spoken punctuation everywhere, and @mentions and slash commands in chat.
  • A dictated message waits 3 seconds before sending. Say “send it” to skip the wait.

Push-to-talk

Choose a key in your settings for this browser: Right Alt Right Ctrl F8 or `. Hold it in any field, release to stop. It stays off until you pick one, so nothing listens unless you ask.

Say what to change

Select text anywhere, press Alt Shift R, and say “make this clearer”. Editable text is rewritten in place. Anywhere else, the new version appears next to it.

Or ask in one phrase

After a draft: shorter, longer, more formal, more casual, friendlier, a bulleted list, or fix the grammar. Or say it your own way.

Voice commands

Get around without reaching for the mouse.

Press Alt Shift V and say where to go or what to start. A new ticket opens with its title filled in and waits for you to check it. Nothing is filed without you.

  • Open #deploysJumps to the channel.
  • Create a ticket for the login timeoutOpens a new ticket in the current project with that title.
  • Search for billing retriesOpens search with the words filled in.
  • Go to the knowledge baseAlso tickets, meetings, chat, settings and home.
Talk with your agents

Ask out loud. Hear the answer.

Your agents already know your tickets, docs and conversations. Voice turns that into a conversation you can have with your hands on something else.

Replies read aloud

Turn on Speak and agent replies are read as they arrive, sentence by sentence. Code and links are skipped. Give each agent its own voice and speaking speed in its settings, and it sounds the same wherever it answers.

Hands-free conversation

Start a voice conversation and take turns. A tone tells you when it is listening, replies stay short, and saying “stop” ends it. Start talking and the agent stops to listen.

Live voice

A real-time spoken conversation with an agent, in sessions of up to 10 minutes. Everything said is written into the conversation, marked as said live. Available to workspaces that allow US processing.

Voice notes agents understand

Record up to 5 minutes, or attach an audio file. Agents hear it before they answer, and the transcript sits under the player so you can read it instead. Voice notes sent on Telegram and WhatsApp work the same way.

Answered in your messaging app

Send a voice note on Telegram or WhatsApp and the agent answers there. Switch spoken replies on for that connection and the answer comes back as a voice note too.

Speak a ticket

Describe the work out loud in the new-ticket form. Type, priority, due date, tags and assignee are filled in for you to review.

Take back what you said

On any dictated message: restore exactly what you said, tidy it up again, or report a wrong transcript.

Phone

A number that answers and writes it down.

Point a phone number at your workspace and calls stop going to voicemail nobody plays. The call is answered, the caller leaves a message, and the recording is transcribed and posted into a channel with a short summary.

  • Connect Twilio and point the number’s voice webhook at the address shown on your phone settings page.
  • Every call lands in the channel you choose, with the transcript and the summary together.
  • Your team reads it, replies in the channel, and turns it into a ticket like any other message.

What it does

Answers the call, takes the message, transcribes it and summarises it where your team is already working.

What it does not do

It is not a live conversation. The caller is not talking to the agent, and nothing the caller asks for is acted on. The call is taken down and summarised, and a person decides what happens next.

If you want to talk to an agent

Use live voice in the browser, or send a voice note from Telegram or WhatsApp.

Meetings

Meetings that end with tickets, not notes.

Record in the browser or upload a file up to 100 MB. No bot joins your call. The meeting page plays the recording back beside the transcript.

1. Record

Start recording from the Meetings page, or upload audio or video.

2. Who said what

Shorter meetings are split into speakers, worked out from the conversation. Rename a speaker once and every line updates.

3. Summary

A short summary and up to 20 action items, written when the recording is processed.

4. Tickets

Turn any action item into a ticket in one click, linked back to the meeting.

Your words

It learns how your team talks.

Product names, customer names and acronyms, spelled the way you spell them, and spoken the way you say them.

  • A shared word list of up to 200 entries: “juju” is written as JuhJuh.
  • Import and export it as CSV to share it between workspaces.
  • Each “Wrong transcript” report suggests words for admins to add in one click.
  • Fix the same word twice and you are offered to remember it.
  • Up to 100 pronunciations, so spoken replies say JuhJuh as “joo-joo”.

Heard “cooper netties”

Kubernetes

Heard “juju”

JuhJuh

Written “SLA”

Spoken “S. L. A.”

You stay in control

Listening only when you ask.

No wake word and no always-on microphone.

You start it

The microphone opens on a click, a held key or a conversation you started, and closes when you stop.

Private meetings

A recording is visible to the person who made it and to workspace admins, and deleting it removes the audio. Audio is deleted 30 days after the recording. The transcript and the summary stay.

Workspace rules apply

Voice notes, meetings, phone messages and live voice follow your plan and the data regions your workspace allows. Dictation uses your browser’s built-in speech recognition.

Questions

Common questions about voice

Does JuhJuh listen all the time?

No. The microphone opens only when you press the mic button, hold your push-to-talk key, or start a voice conversation. There is no wake word, and push-to-talk stays off until you choose a key.

What happens when I correct myself mid-sentence?

Say "scratch that" to drop the last clause, or "never mind" to clear the draft. Ending on "no wait" or "I mean" keeps the microphone open so you can finish the thought, and clean-up applies the correction before anything is sent.

Can I stop a dictated message from sending?

Yes. A dictated message waits three seconds before it sends. Press Keep editing or start speaking again and it stays in the box. Saying "send it" sends straight away.

How does it learn our product names?

Admins keep a workspace word list of up to 200 entries, imported or exported as CSV. Each Wrong transcript report suggests new entries, and a pronunciation list of up to 100 entries tells spoken replies how to say names.

Which browsers are supported?

Dictation, voice commands and voice conversation use the speech recognition built into your browser, so they need one that has it, such as Chrome, Edge or Safari. Voice notes and meeting recording work in any modern browser, and meetings can also be uploaded as files.

Can an agent take phone calls?

It takes a message. A call to a connected number is answered, the caller leaves a message, and the recording is transcribed and posted into a channel with a short summary. It is not a live conversation, and the agent does not act on what the caller asks for. Connect Twilio and point the number's voice webhook at the address on your phone settings page.

Can agents answer a voice note with a voice note?

Yes. Send a voice note on Telegram or WhatsApp and the agent hears it and answers there. If spoken replies are switched on for that connection, the answer comes back as a voice note in the voice you gave that agent.

Who can see a meeting recording?

Only the person who recorded it and workspace admins. The meeting page plays it back beside the transcript, and the audio is deleted 30 days after the recording while the transcript and summary stay. Deleting a meeting removes the audio, transcript and summary; tickets already created from it stay.

Where voice goes

Stop typing what you could say.

Dictation, commands, spoken replies and meetings, in the workspace your agents already know.

JuhJuh uses cookies

We use essential cookies to run the site, and optional analytics cookies to count visitors. We set no advertising or marketing cookies. Only essential cookies are turned on by default. Cookie Policy

Essential

Required for the site to function

Analytics

A first-party visitor ID, so we can count returning visitors. Not shared with anyone.