Google Bard goes global with 40+ languages and image prompts

If you tried using Google Bard before today, chances are you hit a geoblock or gave up because it only spoke English.

That changes now. Google just pushed Bard’s biggest expansion to date, rolling out access across Europe and Brazil, while adding support for over 40 languages including Chinese, Arabic, German, Hindi, and Spanish.

Along with new borders, Bard is picking up a suite of practical tools that address common chatbot annoyances.

Hear it speak and dial in the tone

Text on a screen only gets you so far. Starting today, you can hit a sound icon to listen to Bard read its answers aloud across all 40-plus languages. It is handy for checking pronunciation or hearing how a script actually sounds.

Google is also letting you tweak how Bard talks without typing out a whole paragraph of instructions. In English, you can now toggle between five styles: simple, long, short, professional, or casual. If a draft listing for an armchair is too wordy, you just select short from the menu.

Google Lens arrives, plus easier code export

On the workflow side, Bard is integrating Google Lens for English prompts. You can upload an image alongside your text to ask for details, context, or caption ideas directly from the photo.

Organization gets a needed tune-up as well. You can now pin and rename conversations in the sidebar, making it easier to return to past brainstorming sessions.

Sharing chats is now handled through simple shareable links. For developers running Python, Bard adds one-click exports directly to Replit, alongside the existing Google Colab support.

Why it matters

For everyday users, the friction is basically gone. You no longer have to prompt-engineer your way to a shorter answer or copy-paste code snippets into external editors. The addition of spoken answers and visual inputs makes Bard immediately useful as a daily brainstorming partner.

For the AI landscape, bringing Bard to Europe is a massive regulatory milestone. European watchdogs are notoriously strict about privacy, and Google noted it worked directly with regulators before opening the gates. This proves frontier models can expand under strict compliance.

We think Google played this smartly by leaning on existing ecosystem strengths like Google Lens and Replit rather than just hyping raw model benchmarks. The immediate downside is that key features like visual search and tone adjustment are still English-only. Until those capabilities achieve true parity across all 40 languages, the experience outside English remains a bit second-tier.