Most of us spend a lot of time tapping out messages on the phone keyboard. It works, but it’s not exactly natural. Speaking is something humans have been doing for at least 135,000 years, and we’re pretty good at it. The average person speaks about 150 words per minute, but types about 36 words on the phone. Nobody thinks that gap is worth bridging, and today It’s launching a new feature Nothing Essential Voice is said to do just that.
Essential Voice is a voice-to-text tool built directly into the keyboard and accessible via a long press on the Essential key. Speak naturally, and it turns your words into clear, structured writing. This is the main difference from standard dictation. Traditional dictation writes everything down word-for-word, including filler words, false starts, and awkward pauses. Nothing Essential Voice skips all this and delivers full-sounding text instead.
What can you do with it now
The feature launches with four main capabilities. Auto-correct handles clarity and structure, and removes filler words like “um” and “uh.” Personal mapping lets you set custom voice shortcuts for repeated phrases, links, templates, or specific words. It also has a translation agent, which lets you speak in one language and output text in another. Language support includes over 100 options with auto-detection and the ability to choose regional variants such as Latin American Spanish or Simplified Chinese.
Nothing Essential Voice is being released first on Phone (3) and Phone (4a) Pro, with Phone (4a) following soon after. It expands the extensive Essential AI suite that has been built from scratch since Essential Key was introduced in the Phone (3A) series last year.
Future updates will add context awareness, so the feature can adjust the tone depending on where you’re writing. A message, a work email, and a search field all call for different language, and nothing says it plans to focus on that. The company also mentioned expanding Essential Voice across its broader product lineup as part of a larger effort toward voice-first interactions.
