惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园 - Franky
N
Netflix TechBlog - Medium
宝玉的分享
宝玉的分享
Google DeepMind News
Google DeepMind News
腾讯CDC
G
Google Developers Blog
Martin Fowler
Martin Fowler
Microsoft Security Blog
Microsoft Security Blog
Recent Announcements
Recent Announcements
爱范儿
爱范儿
Engineering at Meta
Engineering at Meta
Microsoft Azure Blog
Microsoft Azure Blog
A
About on SuperTechFans
aimingoo的专栏
aimingoo的专栏
有赞技术团队
有赞技术团队
Jina AI
Jina AI
人人都是产品经理
人人都是产品经理
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
M
MIT News - Artificial intelligence
罗磊的独立博客
博客园 - 三生石上(FineUI控件)
美团技术团队
WordPress大学
WordPress大学
阮一峰的网络日志
阮一峰的网络日志

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
How to Run a Private AI Meeting Notetaker (Zoom and Googl...
Mohammed Ali Chherawalla · 2026-06-25 · via DEV Community

When you invite a cloud notetaker to a call, a bot joins, records everyone, and ships the audio to a vendor's servers. Half the room often does not know it is there. Off Grid AI Desktop is a free, open-source app that records, transcribes, and summarizes your meetings on your own Mac or PC, with the recording in plain sight and the data never leaving the machine.

GitHub →

Free, open-source, runs offline. No account, no API key, no telemetry.

Why a Local Notetaker

A meeting note is supposed to save you time, not hand your conversations to a third party. Here is what changes when the notetaker lives on your desktop.

There is no bot joining the call, so the only consent at stake is yours and your participants', not a vendor's terms of service. The recording, the transcript, and the summary all stay on your disk. There is no per-minute fee, so it costs nothing to keep notes on every call you take.

What You Need

The notetaker runs Whisper for transcription and a local LLM for the summary. Both want a little memory.

Tier macOS Windows RAM / VRAM Free disk
Minimum Apple Silicon M1 Any modern CPU or iGPU 8 GB 10 GB
Recommended M2 / M3 / M4 NVIDIA RTX (6 GB+ VRAM) 16 GB+ 20 GB

The minimum tier handles a smaller Whisper model and a compact LLM fine. The recommended tier runs larger models and finishes the transcript faster. Both macOS and Windows are supported.

Consent Comes First

The recorder does nothing until you press start. While it runs, a visible recording indicator stays on, so anyone watching your screen knows a capture is in progress, and you stop it with one action. There is no always-on listening and no bot lurking in the participant list.

That design is the point. A notetaker should make recording obvious, not hide it. You decide which calls get captured, and the indicator makes it honest for everyone in the room.

How It Works

The recorder captures the screen video, the system audio, and your microphone together, so both sides of a Zoom or Google Meet call are covered. When you stop, the bundled whisper.cpp engine transcribes the audio locally, with no upload.

Then the local LLM turns the transcript into a real note: a title, a summary, and the people who came up. You get the outcome of the call without rewatching it.

The summary does not just sit in a folder. It folds into your private memory and your Day, the persisted journal of what you did, so a meeting becomes part of a searchable record of your work instead of a one-off file you forget about.

Privacy: Stronger Than Otter or Fireflies

Cloud notetakers like Otter and Fireflies route your meeting audio and video through their servers. The recording, transcript, and summary live in their account, under their retention policy, with their access. Off Grid AI Desktop keeps all of it on your machine, because nothing in the pipeline touches a server.

It is AGPL-3.0 licensed, so the code is auditable rather than a black box. There is no account to breach and no telemetry phoning home. On an air-gapped machine, every part still works.

Cloud notetaker Off Grid AI Desktop
Where audio goes Vendor's servers Your disk
Bot in the call Yes No
Per-minute fee Often None
Code you can audit No Yes (AGPL-3.0)
Works offline No Yes

Getting Started

  1. Download or clone from the GitHub repo.
  2. Install and launch the app on your Mac or PC.
  3. Open the Models browser and download a Whisper model plus a local LLM.
  4. Press start before your call, confirm the recording indicator is on.
  5. Stop when the call ends, then read the note in your Day.
git clone https://github.com/off-grid-ai/desktop
cd desktop
npm install
npm run dev

What's Coming

  • Cross-device sync so meeting notes follow you between machines.
  • Unified search across meeting summaries and the rest of your memory.
  • More transcription and summary models as new open-weight releases ship.

FAQ

Q: Is it really free?

Yes. The app is free and open-source under AGPL-3.0. No subscription and no per-minute charge.

Q: How is this different from Otter or Fireflies?

No bot joins your call, and nothing uploads. Recording, transcription, and the summary all happen on your machine.

Q: Does it work with Zoom and Google Meet?

Yes. It records the screen, system audio, and mic, so it captures both sides of any call.

Q: How does it handle consent?

Recording is explicit start-and-stop, and a visible indicator stays on the whole time. Nothing records silently.

Q: Does it work offline?

Yes. Recording, transcription, and summarizing all run with no network connection.

Q: Where do my meeting notes live?

On your disk, folded into your private memory and your Day. No vendor account holds them.

Keep your meeting notes private, on hardware you already own. GitHub →