Google AI Edge Foresight: offline meeting notes, no cloud copy
Google AI Edge Foresight is an experimental Mac note-taker that runs fully offline on Gemma 4 and EmbeddingGemma 2. Why local AI notes matter for client calls.
Google released Google AI Edge Foresight on October 8, a Mac app that takes meeting notes and, per Google's FAQ, can work completely offline. The AI runs on your laptop, not on a vendor's server. For anyone who has taken a pricing call, a hiring call or a legal call with a cloud notetaker in the room, that is the point.
What actually happened
Per TechCrunch and Google's product page:
- Platform: Mac, optimized for Apple Silicon. It works offline, including for in-person meetings.
- Models: EmbeddingGemma 2, at 740 million parameters, handles retrieval on the device. A Gemma 4 model powers the assistant that answers your questions.
- Workflow: A split screen shows your shorthand on one side and AI-written notes on the other. You can view the transcript and chat with the assistant.
- Knowledge base: You can add PDFs, Google Docs, Office files, plain text, Markdown and web bookmarks. When the meeting touches that material, it can answer questions in real time.
- Status: Notebookcheck describes it as experimental, a showcase for Google's on-device models.
Google has not published pricing or accuracy benchmarks for the app. We found no independent accuracy tests yet.
Why local meeting notes matter for your business
A cloud notetaker is a database of your unguarded speech. We covered what happens when that database is missing one access rule. Local notes remove the vendor's server from the threat model.
Your client contracts may already require it. NDAs, health data and legal work often forbid third-party processing. An offline note-taker is easier to defend in a client's security questionnaire.
The pattern is the product. A small embedding model plus a small local LLM over your own files is a stack you can build today. The same pattern works for a quoting assistant or a support lookup tool.
Experimental means do not standardize on it. Use it to test local AI on real calls. Keep your current process until accuracy and support are clear.
Key takeaways
- Google AI Edge Foresight takes meeting notes on a Mac and can run fully offline
- EmbeddingGemma 2 handles retrieval; Gemma 4 runs the assistant, both on the device
- It reads your PDFs, Docs, Office files and bookmarks as a knowledge base
- It is experimental, with no published pricing or independent accuracy tests
- Local AI notes keep sensitive client calls off third-party servers
Need AI on data that cannot leave the building? We build local-first assistants on open models and your own files, so nothing goes to a vendor you did not choose. See how we build it or tell us what has to stay private.
Sources: TechCrunch, Google, Notebookcheck.
- #ai-notetaker
- #on-device-ai
- #gemma
- #privacy
- #meeting-notes
Tommy Rush — Founder, Rush Commerce
Operator turned builder. 15+ years running operations — now shipping the systems businesses run on. More
Get The Rush Report weekly — one email, zero fluff.
Keep reading
Surface RTX Spark Dev Box $5,999: price local AI first
Microsoft's Surface RTX Spark Dev Box costs $5,999 and the Surface Laptop Ultra starts at $2,599. Price local AI against your API bill before you preorder.
Read itClaude for Google Workspace beta edits Docs, Sheets and Slides
Anthropic's Claude for Google Workspace beta puts Claude in Docs, Sheets and Slides on all paid plans. The add-on is scoped; the connectors follow your Drive sharing.
Read it