Google's next Gboard upgrade turns your rambling into something you can actually send

I’ve been covering Android since 2023, when I joined Android Police, mostly focusing on AI and everything around Pixel and Galaxy phones. I’ve got a bachelor’s in IT with a major in AI, so I naturally view technology differently. I usually take a pro-consumer angle instead of the marketing hype, which you’ve probably noticed in my writing.

Humans talk about three times faster than they type, but we don't want to sound like malfunctioning robots while doing it.

By creating an account, you agree to our Terms of Use and Privacy Policy.

You also agree to receive our newsletters; you can unsubscribe any time.

*Required: 8 chars, 1 capital letter, 1 number

By creating an account, you agree to our Terms of Use and Privacy Policy.

You also agree to receive our newsletters; you can unsubscribe any time.

The company introduced Gemini Intelligence as the new foundation for its mobile OS, and one standout feature inside it was a Gboard upgrade called Rambler.

Old dictation makes you talk like a news anchor reading a teleprompter by planning each sentence in your head before tapping the mic.

Rambler is built for the messy and unscripted way people actually speak and tries to solve the part of speech-to-text that usually breaks the flow.

Gboard's AI is often overlooked, but it's a powerful tool we use daily without realizing it

Legacy voice typing types exactly what it hears, with no sense of context or intent, and captures every awkward pause and filler.

Powered by Gemini, Rambler listens to your natural speech and filters the noise in real-time.

If your speech is padded with habitual "ums" and "ahs," Rambler ignores them and keeps the flow. It also understands mid-sentence corrections.

You can say, "Let's meet at 1 p.m., actually, never mind make that 3 p.m.," and instead of dumping that fumble into the message box, it reads your change of intent and writes the corrected time.

As a bilingual user, I have always had problems with native voice typing. Switching languages has improved in modern Gboard, but it can still be limited by device and language.

Rambler pushes that further by using Gemini's multilingual model to handle code-switching inside a single message.

Try mixing languages in one breath, and standard dictation hands you a mangled phonetic mess. Google has built Rambler for people who code-switch every day.

The multilingual model lets speakers move between languages inside a single sentence without losing the thread.

All that real-time conversation magic needs serious computing power, so Rambler is skipping many budget Android phones.

It's scheduled for a summer 2026 release on premium Android phones like the Google Pixel 10 and the Samsung Galaxy S26 series, which have the silicon to run Gemini Intelligence locally.

Still, any tool that listens to your raw and unfiltered thoughts is going to raise privacy concerns. After all, voice is personal data and, in many ways, biometric data too.

Google says it has leaned on a privacy-first architecture. Gboard will show a visual indicator whenever Rambler is working.

Google also said the audio is processed for transcription and that clips are not stored or saved.

Still, to get the benefits of real-time AI editing, you have to trust that Google is actually holding to that promise.

This striking-looking addition to the Pixel line offers a slew of Gemini features, an 5x telephoto lens, and seven years of updates, making this a smartphone that will last you a while.

The Galaxy S26 is Samsung's base flagship. It features a 6.3-inch Dynamic AMOLED display with a 1080p resolution. It includes the Snapdragon 8 Elite Gen 5 chipset and 12GB of RAM. It ships with One UI 8.5 installed, and Samsung promises seven years of software support. In addition to the increased screen size, the S26 also features a larger 4,300mAh battery for 2026.

Rambler is a win for the average Android user, but a nightmare for software startups because Google is about to bake it into Gboard for free.

The dictation market has exploded with AI transcription apps like Wispr Flow and Typeless. These apps charge a monthly subscription for more or less the same result.

Built into the default keyboard

Third-party tools also add complications since they push users into difficult overlays or full keyboard replacements to work.

After premium AI dictation becomes the default, it gets much harder to convince people to install and pay for a separate app. The only real exception is if their phone cannot run it.

Google had already tested the concept through Google AI Edge Eloquent, an iOS-only experiment. (Really letting the Android crowd feel loved here.)

That app showed Google was experimenting with offline-first voice cleanup using on-device Gemma models.

Now, a similar cleanup idea is going into Android's default keyboard experience through Rambler.

If it holds up outside Google's controlled demos, it might finally close the gap between how fast we think and how fast we can get the words down.