Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Google brings Dreambeans to India, its first market outside the U.S.

Google expanded its experimental Dreambeans app to India on Monday, its first market outside the United States, with support for 13 local languages. The prompt-free app turns data from connected Google services into personalized illustrated stories, making India the first real test of whether the format travels.

Published
谷歌 Dreambeans 登陆印度,为首个美国以外市场
Image source: techcrunch.com

Google brought Dreambeans, its experimental AI app, to India on Monday, making the country the first market outside the United States for a product that turns a user's digital activity into personalized illustrated stories. The company is betting on the world's second-largest internet market to widen the audience for that format.

The app is now available on Android and iOS in India with support for 13 local languages, including Hindi, Punjabi, Tamil, Telugu and Bengali, alongside English. The expansion lands less than two months after Google opened Dreambeans to all eligible U.S. users aged 18 and older.

Dreambeans debuted in June and works differently from a conventional chatbot. Rather than waiting for a prompt, it draws on connected Google services — Calendar, Gemini, Gmail, Photos, Search and YouTube — to generate stories about a user's interests and everyday life. Illustrations started out static; Google recently added a feature that uses its Veo video model to animate images in select stories.

Google product manager Gozde Oznur said India was chosen because of strong user interest and an unusually high overlap of people using Gemini, Gmail and Google Photos together, which gives the app more material to personalize from.

Interest has not translated into scale yet. According to Sensor Tower estimates shared with TechCrunch, Dreambeans has recorded roughly 55,000 U.S. downloads since June, about 63% of them in September, when Google opened the app to all eligible U.S. users. Downloads then fell 84% month over month to about 5,500 in October. By comparison, Meta's Muse logged an estimated 7.4 million downloads in its first 30 days, and ChatGPT 2.8 million.

Language support is both the pitch and the limitation. Users can select only one language at a time for their generated stories; the app can read information across Google services in multiple languages but cannot currently mix them. That may narrow its appeal in a country where many people routinely switch between Hindi and English, often blending the two into Hinglish.

Privacy questions follow the product design. Dreambeans relies on Google's Personal Intelligence technology to read connected services that may hold sensitive personal information. Google says users choose which services to link, can disconnect them at any time and can delete their Dreambeans data in the app, with deletions from the underlying services also removed from the personalization system. The app is limited to personal Gmail accounts, and stories are visible only to the account holder.

Proactive generation carries an emotional risk as well. Oznur acknowledged the app could produce stories involving people who have died, and said users can ask it to avoid certain subjects, similar to controls in Google Photos that limit which memories surface. She added that Google is exploring ways to make Dreambeans more actionable, moving from stories and suggestions toward helping users act on them.

For Google, India is a test of whether personalized AI can scale beyond early adopters: language coverage is the entry ticket, but the harder question is whether the product fits multilingual daily life while keeping enough trust around personal data.

Why it matters

The launch tests whether proactive, prompt-free personalization works outside the U.S. and puts Google's data-access and emotional-boundary choices in front of a far larger multilingual audience.

GoogleDreambeansGemini
Back to realtime news

Nearby Updates

All

11/03, 11:40

Ebix and intentyx launch behavioral monitoring tools for AI agents

Ebix and intentyx have launched behavioral monitoring tools aimed at AI agents, according to a Trend Hunter report, placing the offering in the fast-forming AI agent security category. As enterprises hand more work to agents that call tools and touch data on their own, oversight of what those agents actually do is becoming a deployment prerequisite.

11/03, 10:59

Anthropic reportedly plans to offer IPO shares to Australian retail investors

Anthropic reportedly plans to reserve part of its IPO for retail investors in Australia, according to 36Kr, which cited unnamed sources and offered no size, pricing or timetable. The company has not confirmed the arrangement, leaving the allocation detail an unverified signal about how the AI developer intends to widen its shareholder base.

11/03, 12:04

OpenAI Partners With Security Firms to Bolster AI-Powered Cyber Defence

OpenAI is partnering with security firms to strengthen AI-powered cyber defence, ANI News reports, pairing its models with established threat intelligence and incident-response expertise. The report names no partners and discloses no terms, but the move points to frontier labs embedding their models deeper in industry operations.

11/03, 12:44

Anthropic releases FlashProve for cryptographic AI inference proofs

Anthropic has released a product called FlashProve, positioned as a way to produce cryptographic proofs for AI inference, according to Unite.AI. The release pulls verifiable-computing research toward shipping product, with potential consequences for how model vendors and customers allocate trust and responsibility.