Google expands Gemini AI with Notebooks and voice-driven workflows

Google is expanding the capabilities of its Gemini AI application through new organizational and voice-driven features. The company has introduced Notebooks, a feature that allows users to organize conversations, files, and research sources into dedicated workspaces. This update includes bidirectional synchronization with NotebookLM, enabling real-time syncing of sources and chats between the two platforms.

Subscription tiers dictate usage limits for Notebooks: Google AI Ultra subscribers can utilize up to 600 sources per notebook, while free users are limited to 50 sources. This move places Google in direct competition with project-based organization tools from OpenAI and Anthropic.

Additionally, Google announced Gemini 3.5 Transcribe, a speech-to-text model designed for macOS and Android. This model moves beyond simple dictation by using voice commands and screen context to perform complex tasks. On macOS, users can use voice to summarize local files, repurpose text across different applications, and generate images at the cursor location by combining spoken instructions with visible screen content.