Google releases AI Edge Foresight, an offline-capable Mac app for meeting notes that works with on-device models and documents.
Google unveils AI Edge Foresight, a Mac app that runs offline using on-device AI to transcribe meetings and generate notes. It also lets you upload documents to build a personal knowledge base and asks questions in real time using the embedded Gemma model. The app is optimized for Apple Silicon and can operate without internet access, making it handy for travel or areas with poor connectivity. A consumer release date and pricing aren’t yet announced.
Why it matters: offline AI tools protect privacy and work where networks aren’t reliable. For everyday users, it could speed up meeting notes and help find information quickly across notes and documents.
Google has released a new local-first AI note-taker called AI Edge Foresight for Mac, designed to run entirely on-device and offline. The app uses Google’s EmbeddingGemma 2 model, a 740-million-parameter on-device model, to capture meeting notes across apps without needing a constant internet connection.
Like popular note-takers such as Granola, AI Edge Foresight offers a split-screen setup. One side lets you jot shorthand notes while the other shows AI-generated notes or summaries. You can also keep manually written notes while transcripts appear in real time, providing a flexible workflow for people who attend meetings in various apps or in person.
Beyond transcripts, the app lets you drop in a range of documents and links to build a personal knowledge base. This includes PDFs, Google Docs, Microsoft Office documents, plain text files, Markdown, and bookmarked web pages, so the assistant can pull relevant details and answer questions from your notes and sources as you work.
Google emphasizes that AI Edge Foresight is optimized for Apple Silicon devices and can operate completely without an internet connection. This makes it appealing for travelers, people in areas with spotty connectivity, or anyone who wants to keep sensitive notes on-device rather than in the cloud.
The release continues Google’s broader push into collaborative AI tools that work offline and across apps. While Google previously offered an experimental dictation tool that could run offline, this new product is positioned as a more complete meeting-notetaking solution that could hint at consumer-facing features across Gemini apps in the future, depending on market reception.
AI Edge Foresight is a Mac app designed for offline use with on-device models. It transcribes meetings, supports manual note-taking, and lets you view and converse with a Gemma 4-powered assistant to get answers related to your notes and uploaded documents.
Google’s product is described as a competitor to Granola, offering a similar split-screen note-taking experience and real-time transcription. The key difference highlighted is its on-device operation and the ability to build a knowledge base from uploaded documents to answer questions during or after meetings.
For people who shuffle between video calls and in-person meetings, AI Edge Foresight promises a more private workflow since data can stay on your device. It could also speed up note-taking, reduce context-switching, and provide quick access to information buried in documents when you need it during a discussion.
If you often work offline or travel, this could be a practical option because you won’t rely on cloud connectivity to capture and retrieve your notes. As with any AI tool, you’ll still want to review transcripts for accuracy, especially when conversations include nuanced or sensitive details.
Google hasn’t announced a consumer-wide rollout date or pricing yet. The company’s emphasis on Apple Silicon optimization suggests a Mac-first approach before considering other platforms. If the early performance holds up, we could see broader adoption among professionals who rely on note-taking during meetings and document-heavy workflows.
So far, Google describes it as a Mac app optimized for Apple Silicon, with offline capabilities. No other platforms are confirmed yet.
The app is designed to work offline using on-device models, but users can upload documents to build a local knowledge base. Cloud storage usage isn’t described in the release.
Google hasn’t announced a public release date. Expect more details as the company rolls out information on a potential consumer version and its cross-app integration plans.
You're reading the quick version.
Uber and Pony.ai will start London-based tests of Gen-7 robotaxi tech soon as part of a broader European expansion for autonomous rides.
A new, single prompt-driven helper ties together calendars, emails, documents and coding across tools, with built‑in governance.
Vesta raises $30 million to deploy AI agents across loan origination, promising faster closings and lower labor costs for lenders.
The Daily Brief
Every morning, the news that matters from AI, phones, apps and the people shaping tech. Explained in plain English. Free.
One email a day. No spam, unsubscribe anytime. Privacy policy
