What is Speech Recognition and Synthesis from Google?

Speech Recognition & Synthesis from Google is the speech engine that Google publishes for Android phones. It has two jobs. It turns what you say into text for Google apps and for other apps that ask for it, and it reads text on your screen aloud. You do not usually open it like an ordinary app. You meet it through voice features in other apps and through your phone's settings.

The name can be confusing. Google's Play Store listing uses the title Speech Recognition & Synthesis for the app, and the older name, Speech Services by Google, still appears in some places. Our glossary entry on Speech Services by Google explains that they refer to the same engine.

It turns your speech into text for Google and other apps

Google's listing says the speech-to-text feature is provided to Google and to other third-party apps, so they can convert what you say to text ¹. The listing gives Google Maps voice search and the Recorder app's transcription as examples of apps that use it.

Because several apps share the same engine, a voice problem can come from the engine rather than from the app you are typing in. If dictation works in one app and fails in another, the engine is not the only cause, but it is one thing to check.

It also reads text aloud

The same listing says the app powers applications that read the text on your screen aloud ². This is the text-to-speech half of the engine. It does not type anything for you. For the general terms, see our entry on automatic speech recognition.

You can choose it as the voice input engine

Google's listing tells you to select Speech Recognition and Synthesis from Google as your preferred voice input engine ³. The listing gives this menu path: Settings, then Apps & notifications, then Default apps, then Assist App. It was read on 2026-10-17. Android versions and phone makers rename or move these menus, so if you cannot find them, search your settings for "voice input" or "default apps."

Changing this setting changes which engine handles voice input. If speech-to-text suddenly behaves differently from one app to another, this is the first place to look.

What the listing does not tell you

The listing does not say whether recognition runs on your phone or on Google's servers. It also does not say which keyboards use this engine. Those two answers decide whether dictation works offline and where your audio goes, so they matter more than the name.

Google's own voice typing gives a clue about the direction Google has taken, but it is a different product. In February 2025, 9to5Google reported that Gboard for Android features offline voice typing on all devices ⁴. That report is about Gboard, the keyboard, and it does not tell you what this engine does on your phone. For the keyboard side, see our entry on Gboard and the guide to Google voice typing.

How to check what your phone uses

Use a short test rather than a guess:

1. Open your phone's voice input settings, using the menu path above, and note which engine is selected.
2. Turn on airplane mode, so the phone has no internet connection.
3. In the app you use most, tap the microphone key and dictate one sentence.
4. If the sentence appears, recognition worked without a connection for that app and language, which suggests it ran on the phone. If it does not, the app probably needs a connection, or the offline language data is not installed.
5. Turn airplane mode off and repeat the test with a sentence of the same length, so you can compare the results.

This test shows the behavior of your phone and your keyboard, not the internal design of the engine. The listing still does not say how the engine works.

Can you turn it off?

We did not verify steps for disabling or uninstalling this app. The search results we read on the subject were unreliable, and the listing does not cover it. The documented step is the one above: choosing a different preferred voice input engine changes which engine handles voice input. Menu names for the text-to-speech engine also vary by phone, so we leave out exact steps until a source we can open confirms them.

If you want to stop an app from using voice input, the safer first step is the app's own microphone or voice setting, or your phone's microphone permissions for that app.

Related pages

• The guide to on-device and cloud speech recognition compares where the recognition runs, with the accuracy, privacy and cost trade-offs.
• The speech recognition overview covers how the technology works and how it is measured.
• Does Google have a voice to text app? explains which Google products handle dictation.
• Is on-device speech recognition as accurate as cloud? covers the accuracy question that the engine's location raises.