SpeakribeDictation for WhatsApp
A voice note asks the other person to listen at your pace, and they cannot skim it. Speakribe gets you the speed of talking with the courtesy of text: you speak, they receive a normal message.
Hey, did you get a chance to look at this?
Can you send that report to accounting before end of day?
Why it helps here
Typing on a Mac keyboard is fine until the message is three sentences long, at which point most people send a voice note instead, and voice notes are worse for the recipient, who cannot search, skim, or read them in a meeting.
Text your recipient can actually use
Searchable, skimmable, quotable, and readable in a meeting. Everything a voice note is not.
Hebrew, English, and switching between them
Over 100 languages with no mode switch, which matters most in exactly this kind of casual, code-switching conversation.
Private by construction
Personal messages are transcribed on your Mac and never uploaded anywhere.
Things people say into WhatsApp
- “Can you send that report to accounting before end of day?”
- “Running about ten minutes late, order without me and I'll catch up.”
- “That works. Let's do Sunday at the usual place, and I'll bring the charger I borrowed.”
- “Quick update: the delivery moved to Thursday, so no need to be home tomorrow.”
Setting up for WhatsApp
WhatsApp on a Mac is either the desktop app or web.whatsapp.com, and Speakribe treats them the same because it types into the focused field and does not know which one you are using. The one thing worth configuring here is language.
Set your language, or leave it and switch freely
There is a language setting, but the reason most people leave it alone is that switching mid-sentence works without it. If you write to family in Hebrew and colleagues in English from the same app, this is the setting you do not need to touch.
Click the message box, not the chat
Selecting a conversation in the sidebar leaves focus on the list. Click into the message field itself before speaking, or the text goes to WhatsApp's own keyboard navigation and vanishes.
Keep auto-submit off for group chats
The setting that presses Enter when transcription finishes turns a dictation into a sent message with no gap in between. In a one-to-one chat that is convenient. In a group of forty it removes the two seconds in which you would have noticed the wrong tone.
Add the names you actually message
Family names, place names, the transliterated words that sit between languages. These are the reliable misses in personal messaging, and the custom-words list in settings fixes them permanently.
One dictation, start to finish
A reply that would otherwise have been a ninety-second voice note, spoken in about twenty seconds and delivered as text the other person can skim on a train.
Sorry, only seeing this now. Yes to Thursday, but I can't do before seven because I've got the thing with the accountant at five and it always runs over. If you want to eat first, book somewhere near you and I'll come to you. Otherwise let's just meet at the place from last time and figure it out there.
Sorry, only seeing this now. Yes to Thursday, but I can't do before seven because I've got the thing with the accountant at five and it always runs over. If you want to eat first, book somewhere near you and I'll come to you. Otherwise let's just meet at the place from last time and figure it out there.
- Four sentences from one take. Casual speech produces short sentences, which is why dictated messages read more like messages than dictated documents read like documents.
- "seven" and "five" came out as words, not numerals. In a message this reads fine; if you want the digits it is a manual edit, because nothing infers your preference.
- No emoji, and there is no way to say one. Everything about the message except the words is still typed, which is worth knowing before you replace voice notes entirely.
What it won't do in WhatsApp
No emoji, stickers or attachments
Speakribe produces words. Saying "thumbs up" types the phrase, not the emoji, and nothing about attaching a photo or sending a voice note is voice-driven.
It cannot pick the conversation
You still open the chat yourself. There is no way to say who a message is for, which is arguably the feature that keeps this safe.
It writes text, so it loses what a voice note carries
Tone, warmth and the sound of someone laughing do not survive transcription. For logistics, text is better for the recipient. For the message where how you said it was the point, send the voice note.
Questions people actually ask
- Is this the same as sending a voice note?
- No, the opposite. You speak, and your contact receives ordinary text they can read and search.
- Does it work in WhatsApp Web and the Mac app?
- Yes, both, since Speakribe types into whichever field has focus.
- Can I dictate in Hebrew?
- Yes, and you can mix Hebrew and English in one message without changing any setting.
- Can I mix Hebrew and English in one message?
- Yes, and without changing a setting. Switching language mid-sentence is handled, which is the specific case that makes a single-language dictation tool unusable for bilingual messaging.
- Can it send the message for me when I stop talking?
- There is an auto-submit setting that does exactly that, off by default. It suits one-to-one chats. In a group we would leave it off, because an unreviewed message and a large audience is a bad combination.
Speakribe in other apps
Ready to stop typing?
One payment. Every future update included.