Speak your request with voice typing
In shortClick the microphone in the prompt box (or press ⌘⇧Space on a Mac, Ctrl+Shift+Space on Windows) and talk. Your words appear as you speak, in English or Japanese, with filler words like "um" left out. Click again to stop, or press Enter to finish and send in one go. Hold the shortcut to talk only while it is held.
Describing an app out loud is often faster — and more detailed — than typing it. Voice typing turns what you say into text in the prompt box as you say it, word by word, so you can see it working while you talk. It understands English and Japanese, even mixed in one sentence, and leaves out filler words like um and えーと.
It works in the prompt box on Projects and in every project’s chat.
How do I use voice typing?
Section titled “How do I use voice typing?”-
Click the microphone. On Projects it is in the top-right corner of the prompt box; in a project it is beside Send.
-
The first time, your browser asks to use the microphone — choose Allow.
-
Talk. The button turns red, its bars follow your voice and the box gets a red edge. Your words appear where the cursor was, as you say them.
-
Finish in whichever way suits you:
- click the microphone again (or press Esc) to stop and edit the text, or
- press Enter to stop and send straight away — the last words are added first.
You can start anywhere with the keyboard: ⌘⇧Space on a Mac, Ctrl+Shift+Space on Windows and Linux. Press it once to start and again to stop — or hold it while you talk and let go when you are done, like a walkie-talkie. Holding the microphone button works the same way.
Where do the words go?
Section titled “Where do the words go?”Where your cursor is. Click inside what you have already written and talk to add a sentence there; select some text and talk to replace it. If you had not clicked in the box, the words go at the end.
Spacing and capitals are taken care of: a sentence you continue carries on in lower case, and English and Japanese are spaced the way they are written.
While you talk the box shows your words but you cannot type in it. As soon as you stop, the text is yours to edit — and ⌘Z (Mac) or Ctrl+Z (Windows) takes the whole dictation back in one step.
Which languages does it understand?
Section titled “Which languages does it understand?”English and Japanese, whichever language MonstarX is set to — you can switch between them mid-sentence (“Stripe で決済できるようにして”). Product names like Stripe, Supabase, Figma and GitHub are recognised, and so is your project’s name.
Words that are only hesitations — um, uh, えーと, あの、 — are left out. Words that carry meaning are kept.
Frequently asked questions
Section titled “Frequently asked questions”Does voice typing cost credits?
No. It is included with every plan and uses no build credits. It appears in Usage as _Voice typing (included)_.
Is my voice recorded or kept?
No. Your microphone goes straight to the speech-to-text service while you are dictating and nothing is kept by MonstarX — only the text in your box. The microphone turns off when you stop, when you switch to another tab, or after 45 seconds of silence.
The first words were missing — why?
They should not be: MonstarX holds on to what you say while it connects (about a second) and sends it as soon as it can. If words are missing, check that the right microphone is selected in your browser or computer settings.
It says it needs my microphone.
Your browser blocked the microphone for this site. Click the lock or settings icon beside the address, allow the microphone, then click the microphone button again.
Can I use it on my phone?
Yes. Tap the microphone, talk, and tap it again to stop. On a phone Enter makes a new line, so tap Send to send.
How long can I talk for?
Up to 10 minutes at a time. Voice typing also stops by itself after 45 seconds without speech.