Google AI Edge Foresight for Mac turns scribbles into full notes using offline recordings

Google has unveiled EmbeddingGemma 2, a new AI model capable of organizing, searching, and connecting information natively on consumer hardware. To demonstrate its potential, Google developed an AI Edge Foresight Mac app designed for intelligent note-taking. EmbeddingGemma 2 excels in finding specific video clips from voice memos or searching through lengthy audio recordings using text queries, all processed by a single, multimodal model.

The 740 million parameter model, built on the Gemma 4 architecture, is available under the Apache 2.0 license and designed to run locally, even with limited resources. Quantization allows the model to run on devices like the Google Pixel 11 Pro, requiring only about 191MB of active RAM for text-only weights and approximately 567MB for the full multimodal model.

The AI Edge Foresight Mac app showcases this technology by listening in the background during video or in-person meetings, allowing users to jot down quick shorthand notes. Leveraging EmbeddingGemma 2, the app then automatically transforms these notes into polished, comprehensive notes based on the meeting transcript. All processing occurs on your device, with no need for cloud involvement.

Additionally, the app can integrate with project folders, reference materials, diagrams, and calendars, enabling users to query the system and retrieve relevant information quickly. The app also features a chat function that offers conversational retrieval and summarization of personal knowledge, along with a Live Assistance capability that listens and answers questions in real-time.

Users can download Google AI Edge Foresight here. This joins Google's other Mac apps in the suite, including AI Edge Gallery and AI Edge Eloquent, which focuses on offline transcription.