Google today announced its new EmbeddingGemma 2 model that can be used to “organize, search, and connect information directly on consumer hardware.” To show off its potential, Google created an AI Edge Foresight Mac app for “intelligent notetaking.”

EmbeddingGemma 2 “can help find a specific video clip from a voice memo, or search through hours of audio recordings based on a text query, all processed by a single, natively multimodal model.”

The 740 million parameter model is built on the Gemma 4 architecture and available with an Apache 2.0 license. This is meant to run locally “within tight resource constraints”:

With quantization, on a Google Pixel 11 Pro, EmbeddingGemma 2 requires as little as ~191MB active RAM for text-only weights and ~567MB for the full multimodal model.

To showcase, there are demos in Google AI Edge Gallery on Android and iOS. The bigger showcase is a new Mac app called Google AI Edge Foresight.

During video or in-person meetings, this notetaking app listens in the background so that you can just “scribble shorthand bullets for things you want to remember.”

The app, leveraging EmbeddingGemma 2, “enriches [that shorthand] into polished notes using the meeting transcript instantly.” The transcript and audio are processed entirely on your device rather than the cloud.

Another capability lets you “point Google AI Edge Foresight to your project folders, reference materials, diagrams, and calendar.” Supporting local files and Google Drive, this in turn lets you ask the app questions and quickly get back information.

It automatically understands your work without needing to manually organize your files, to ensure critical information is ready whenever you need it.

Source link