Breaking it down simply — the Nothing Phone (4a), Phone (4a) Pro, and Phone (3) have received a new update that brings a range of improvements and quality-of-life upgrades. The highlight is the Essential Voice feature, which enables hands-free typing for messages, emails, notes, and even searches. Built on speech-to-text, it goes a step further by refining clarity and structure, filtering out filler words, translating input, and delivering cleaner, more polished text.

We received early access to the update and got hands-on with the Essential Voice feature on our Nothing Phone (3). Here are our initial impressions, what it can do and how it improves upon the standard speech-to-text features found on most smartphones.

What makes Essential Voice feature essential?

How to use the Essential Voice feature?

How does the Essential Voice feature work?

Nothing Essential Voice vs standard speech-to-text

Our impressions

Nothing Phone 4a Pro Images

Nothing Phone 4a Images

Nothing Phone 3 Images

What is Essential Voice?

Essential Voice is the latest major edition of Nothing OS and part of the company’s proprietary tool, Essential Space. The feature is powered by AI and available system-wide, on any keyboard across apps, including third-party ones. The Essential Voice transcribe your speech to text, similar to Gboard’s and several other speech-to-text apps. That said, Nothing has gone a step further by formalising your speech-to-text, making it clean and ready to use.

Essential Voice “…will automatically tidy up your words, so your text sounds like finished writing – not a transcript. It’s also built into the places you write, so you don’t have to leave your conversation to benefit,” the company said in its press release.

What makes Essential Voice feature essential?

Nothing lists the various benefits of the Essential Voice feature. The list includes:

Auto-correction: Unlike standard speech-to-text, Essential Voice avoids messy or fragmented output. It refines clarity, removes filler words like “ums” and “uhs,” and delivers cleaner, more professional text.

Personal Mappings: The feature supports custom voice shortcuts for words, links, templates, and frequently used phrases, helping keep your output consistent and error-free.

Translation Agent: It can translate spoken input directly into another language, removing the need to draft in your native language first and then translate separately.

Support for 100+ languages: Essential Voice offers broad language support with auto-detection, along with options to choose regional variants such as Latin American Spanish, French, Portuguese, and Simplified Chinese.

How to use the Essential Voice feature?

To access the newly launched Essential Voice feature, you’ll need a Nothing Phone (4a), Phone (4a) Pro, or Phone (3) – at least for now. It may roll out to older Nothing smartphones, but the company has yet to confirm support or provide a timeline.

If you’re using the latest mid-segment or premium-tier Nothing smartphone, you can check for the update now. To do so, go to the phone’s settings –> System –> System updates. We received the Essential Voice feature on our Nothing Phone (3) with build number Metroid-B4.1-260408-1909-IND.

Download the update and restart your device to start using Essential Voice. One can access it by long-pressing the Essential Key or tapping the small icon in the bottom-left corner of the keyboard.

How does the Essential Voice feature work?

Essential Voice requires access to your smartphone’s microphone and, notably, an active internet connection. It relies on Google Gemini 3 Flash for cloud-based processing. That said, Nothing claims the generated text isn’t stored on its servers and is only sent back to This product, addressing potential privacy concerns. Additionally, Essential Space doesn’t listen in the background and only activates when you trigger it.

Despite not processing on-device, the Essential Voice quickly analyses the speech and generates text, with minimum errors.

Most speech-to-text tools, including Gboard, transcribe speech word-for-word. As a result, the output often includes natural fillers like “uhh” and “umm” that people commonly use while speaking. The Essential Voice feature basically fixes this issue, and here’s how it stands out compared to the standard speech-to-text: Essential Voice (L) vs Gboard speech-to-text (R)

Feature | Essential Voice | Standard speech-to-text
Output | Clean, “ready-to-send” text | Verbatim transcript with ums/ uhs
Grammar | AI-corrected | Raw input
Formatting | Auto-format into lists/ notes | Text block
Access | System-wide via Essential Key | Keyboard-dependent

In our limited time with Essential Voice on the Nothing Phone (3), it proved to be genuinely useful. It consistently delivers accurate results with minimal editing, especially when you speak clearly and articulate your thoughts. There’s also a strong sense of convenience; there’s no need to keep holding a button. Simply tap once to start recording your speech, and tap again to stop. Drafted via Essential Voice (no keyboard was harmed).

Since the feature works across apps, it’s especially useful for drafting emails or for long searches. It feels far more seamless than typing on a smartphone, which can often be tedious on a small screen. That said, by removing filler words, the Essential Voice text can appear a bit formal – something that one may not admire while messaging to a friend or family. Hopefully, a future software update could fix this.

Nothing has promised context-awareness for the feature, which seems to adapt to where you are writing, like messages, work emails, or searches.