Introducing Essential Voice

Zac 🛸 · 2026-04-23T12:00:00+00:00

Hey Community, Right now, the main way we communicate on our phones is through typing. But speaking is more natural and four times faster. That is why voice notes have become so common. But they are not perfect. They take time to listen to, cannot always be played out loud, and are harder to scan when you just want the key point. Traditional dictation has not really solved this either. It turns speech into text, but often leaves you with something rough and fragmented, including all the ums, uhs and stutters you did not want in the first place. That is where Essential Voice comes in. It turns speech into clear, ready-to-send text in real time – the speed and ease of talking with the clarity and flexibility of text. Integrated into the keyboard and the Essential Key, it’s a step towards voice-first interactions with our devices. More than simple transcription Better than a simple transcription, Essential Voice cuts filler words for a more streamlined and considered output. Use it to send messages and emails, or format your speech straight into lists, steps or bullet points to help get your point across. Built for multilingual communication Speak to Essential Voice in more than 100 languages with auto-detection. You can also choose regional variants of languages like English and Spanish. And you can speak in one language and have it translate and transcribe it for you in real time. Shortcuts for the things you say often Think of this like a text shortcut library you build over time to use specific spellings or add additional text without saying it every time. You can say something like “nothing OS”, and a personal mapping would automatically change that to “Nothing OS”. If you have a favourite restaurant, Essential Voice will link an address to the name and type it out in the transcription instantly. Privacy and Data Handling Essential Voice activates only when you choose to use it, and it does not listen in the background. Once activated, your audio recording is encrypted and processed on our server. The generated text is returned to your device. It is not stored on our servers. Availability Available now on Phone (3). Phone (4a) Pro will follow later this month, with Phone (4a) in early May.

Februar92

Zac 🛸 Will the Phone 3(a) Pro also get the feature?

SelvakumarM

Zac 🛸 amazing 🤩 🔥

nothingmore

SelvakumarM

Zac 🛸 own voice assistant give up next generation

Cmf Guy📍

Is it coming to CMF phone 2 pro? It also has essential key like phone 4a series so it should.

arryankeshan

When will this come to the phone 2a ?

DEVRATAN

Please don’t forget to add this feature to the Nothing Phone 2 as well

TarunAbhishek

Zac 🛸 How to use it i am having phone (3)

zurawskki

Zac 🛸 how to use it I don’t have it on my phone 3 🫤✌🏾

Anonymous1465

Underneath the fascade it’s just Google Gemini though…

Thanks Google for this tech I guess? Nothing just wrapping it up and pretending to be an AI company

IceMelter

Are the AI features in the Essential Space and this new Essential Voice run locally on the device? whats the privacy policy/data collection situation on this?

olunar

Will it be available for Phone 3a pro?

Anonymous1465

IceMelter Download a traffic monitor like PCAPDroid and check it - given it’s just Gemini Flash under the hood AFAIK it’s cloud based to some extent anyway

_jQ_o8pDmlOrFp

It’s not working reliably. Sometimes it doesn’t detect the language correctly. Sometimes it simply doesn’t produce results. Sometimes it produces a different text and meaning. It takes a long time to process.

IceMelter

Anonymous1465 yeah that makes sense, is it possible to turn it off?

petar_RvKGbS8Lss

Any idea how to change the language? I see only Language Variants, but only for five languages, English, Spanish, French, Portuguese and Chinese, but no option to add another one. Is there an option to add primary language or something?

AvyuktGupta

What about 3a or 3a pro?

AdolfPupik

The feature works great on my Nothing Phone 3, even in my native language. The feature requires an active internet connection, and according to usage statistics, the AI actually uses Gemini 3 Flash. That’s an excellent LLM model, but it appears that Nothing runs the model itself on its own domain with a server hosted by Amazon. 👍

JustinPemberton

Zac 🛸 This is what I miss most from my Pixel, so I’m so glad it’s on my phone.

RaunakDaga

Hey man will cmf get this update too if yes then when really excited

« Previous Page Next Page »