Direct, spontaneous conversation between people who speak different languages has always required either a skilled interpreter or a reliance on text-based translation tools. Google Meet’s new real-time speech translation changes that equation, rolling out a feature that uses artificial intelligence to translate spoken language live, while preserving the speaker’s unique voice, tone, and emotion.

This update, announced at Google I/O 2025, leverages Google DeepMind’s advanced audio model to make video calls more natural and accessible for users who don’t share a common language. Instead of just reading subtitles or captions, participants now hear a synthetic version of the speaker’s voice in their own language, with original vocal inflections and cadence intact. This approach creates a more authentic, less robotic experience for everyone on the call.

Google Meet's live translation in action. Source: Google

How Google Meet’s Real-Time Speech Translation Works

Google Meet’s speech translation operates directly within the app, requiring no third-party add-ons or separate software. When enabled, the feature listens to the speaker’s words and swiftly generates a translated voice overlay in the listener’s preferred language. The original audio remains faintly audible, but the translated speech—delivered in a synthetic voice closely matching the speaker’s—takes center stage.

Currently, the feature supports English and Spanish, with Italian, German, and Portuguese slated to launch in the coming weeks. According to Google, latency is minimal, so conversations can flow naturally without awkward delays. The technology is powered by a large language audio model, trained to capture not only words but also the subtleties of tone and emotion, allowing translated conversations to feel more genuine.

Eligibility and Access

At launch, real-time speech translation is available in beta to subscribers of Google’s AI Pro and Ultra plans. Only one participant in a call needs to have an eligible subscription for the feature to be activated for everyone. Google is also preparing to expand access to business and enterprise users through Workspace, with early testing planned for later in the year.

To activate the feature, users can select their spoken and preferred languages from a dedicated panel in the Meet interface. Once set, speech translation runs automatically for supported calls, making it straightforward for users to participate in multilingual meetings without complex setup.

Key Benefits and Use Cases

By translating speech in real time and preserving vocal identity, Google Meet’s new feature addresses several longstanding challenges in cross-language communication. The most effective scenarios include:

  • International business meetings, where team members can speak freely in their native languages and still be fully understood.
  • Family conversations between relatives who do not share a common language, such as English-speaking grandchildren and Spanish-speaking grandparents.
  • Educational settings, where teachers and students from different linguistic backgrounds can participate in live discussions.
  • Customer support or client consultations allow service providers to assist users regardless of language barriers.

Compared to traditional live captioning or subtitle translation, this method significantly improves the conversational flow and emotional nuance of video calls. Participants no longer need to read text or wait for delayed responses, resulting in a more immediate and personal experience.


Limitations and Beta Considerations

As with any beta release, users may encounter occasional inaccuracies or brief delays, especially as the system adapts to new languages and accents. Google encourages feedback from early adopters to refine the technology and address edge cases. The company’s roadmap includes expanding language support and integrating the feature more broadly across Workspace and consumer accounts.

For now, speech translation is limited to select subscription tiers and specific language pairs, but Google has stated its intention to make seamless multilingual communication accessible to a wider user base over time.


Google Meet’s real-time speech translation marks a substantial upgrade for anyone who needs to communicate across languages, making conversations more direct, expressive, and inclusive than ever before.