The Gemini app becomes more agentic, delivering proactive, 24/7 help
Key Points
- Gemini Spark: 24/7 cloud agent that executes tasks
- Gemini Omni: multimodal cinematic video generation
- Neural Expressive: redesigned dynamic UI with rich media
Summary
Google updated the Gemini app (May 19, 2026) with agentive features, a redesigned UI, multimodal video generation, and desktop integration. Key releases include Gemini 3.5 Flash (faster frontier model), Neural Expressive (new dynamic UI), Gemini Omni (text/image/video → cinematic video), Daily Brief (personal morning digest agent), and Gemini Spark (a cloud-based 24/7 agent that performs tasks across connected apps). The macOS app is available today; Spark and advanced voice features roll out later this summer.
Key Points
-
Models and platform
- Gemini 3.5 Flash: next-gen model combining high capability with low latency.
- Gemini Omni: multimodal video generation from text, images, and clips; available to Google AI Plus/Pro/Ultra subscribers.
-
UI and interaction
- Neural Expressive: new visual language with fluid animations, richer multimedia responses, integrated live conversation, improved mic and regional dialects; rolling out globally on web, Android, iOS.
-
Agents and automation
- Daily Brief: personalized, skimmable morning digest that aggregates Gmail, Calendar, and other app signals; opt-in and available to AI Plus/Pro/Ultra (U.S. rollout starts first).
- Gemini Spark: cloud-based, always-on agent built on Gemini 3.5 with the Antigravity harness; integrates with Workspace (Gmail, Docs, Slides) and MCP partners (Canva, OpenTable, Instacart) to run background tasks, create workflows, set triggers, and teach skills.
- Safety and control: Spark requires opt-in, explicit app connections, and will ask before high-risk actions (payments, sending emails).
-
Desktop
- macOS app available now; Spark and new voice-driven editing features (local file and cursor-aware actions) coming later this summer.
Rollout and availability
- Neural Expressive: global rollout today (web, Android, iOS).
- Gemini Omni & Daily Brief: rolling out to Google AI Plus/Pro/Ultra subscribers (starting in the U.S.).
- Gemini Spark: trusted testers this week; Beta for U.S. Google AI Ultra subscribers next week; wider features (MCP actions, texting/emailing Spark, custom sub-agents, local browser control) planned over summer.
Engineer notes / practical impact
- Expect new APIs and MCP connectors for third-party integrations (Canva, OpenTable, Instacart initially).
- Spark runs in the cloud but will have desktop integration for local file/voice actions—plan for hybrid auth and local context handling.
- Design teams should review Neural Expressive assets for updated UX patterns and multimedia response capabilities.
- Prioritize privacy/consent flows and safeguards for background agents that can act on inboxes, calendars, and payments.