Getting live
Talk the shop, then read it like an editor.
Create can listen through the browser's speech recognition and dump what you said into the same draft you would have typed. That is useful when your hands are floury or you think faster out loud than on a keyboard. It is not a private better model, not a phone app we ship, and not a way to skip Google sign-in. If the browser has no SpeechRecognition, the studio will say voice input is not available, switch browsers or type. A mumbled paragraph still spends a Cipher turn when you send it, so the job after talking is reading.
Voice input transcribes into the Create draft using the browser's speech recognition. Cipher does not get a special spoken mode. Read the text, fix towns and hours, then send. If your browser cannot listen, type. Spoken turns still count toward fifty on Free.
What the mic actually does
It starts the browser recognizer in your UI language, waits for a result, and appends the transcript to the draft. Interim mumbling is not streamed as a live caption you can edit mid-word. Stopping the mic stops listening; it does not send the brief by itself. You still hit send, which is when a turn is spent.
Speak nouns, not atmosphere
Say the trade, the street, the hours, the from-price, and the action. Do not orate about heritage and passion for ninety seconds. Cipher will garnish a false shop. Spell numbers: open at seven, not "earlyish". If a town name is easy to mis-hear, type that line.
Trade
Towns
Hours
Price posture
Action
Read it like a mean editor
Homophones will betray you: flour versus flower, aisle versus isle. Hours get mangled: "ten to six" might land as "Tennessee" in a bad transcript. Fix those before send so you do not spend a second turn undoing the first. Voice is a first draft, not a publish.
When not to use it
A noisy kitchen will transcribe the mixer. A shared office will put someone else's meeting into your bakery. Legal names, phone numbers, and URLs are safer typed. If recognition is missing, do not fight the mic, paste from notes.
Voice plus photos
Talk the brief, then attach the hero and product shots with the right roles. Do not expect the mic to describe a JPEG. A sentence like "the mill photo is the hero" should be visible in the text you send. Files still have size limits; talking does not raise them.
- Tap the mic
- Say trade, towns, hours, action
- Stop listening
- Edit the transcript
- Attach photos if you have them
- Send once
Privacy and the recognizer
Speech recognition is a browser feature, not a Bizzit-hosted stenographer we document as a SKU. Treat it like talking to the device you are holding. Do not dictate card numbers or other people's private data into a brief. Google sign-in is still how the project is owned after you send.
Limits and caveats
We do not promise accuracy of third-party speech recognition. If the transcript is junk, it is still your message when you send it.
If a sentence would require a competitor SKU we have not verified, or a statistic we did not measure, it does not belong here.
What to do after you read this
If this page answered the query, open Create with the nouns from your own business, not a slogan. If it did not, you are probably in the wrong cluster: go back to the blog hub or Getting live. Do not spin a second URL for a synonym. One intent per page is the whole point of this library.
Tools that actually do this job
These are public pages. Use them when the job is this job. They are not a second product you have to buy.
Questions people actually ask
Which language does voice use?
The recognizer uses the browser language when available. If that is wrong for your town names, type those words.
Does voice skip the fifty-turn pool?
No. Sending the message spends a turn whether you typed or talked.
Can I voice-edit overnight jobs?
You can speak "restock the catalog nightly" into the draft. Pro still owns overnight jobs as a plan line. Speech does not enable them on Free.
Why did the studio say voice is not available?
The browser did not expose SpeechRecognition. Use a browser that does, or type. We do not ship a custom dictation engine.