Donna app icon

Meet the best f*ckingAI assistant in the world

Donna is a private AI assistant that lives on your computer. She remembers your people and projects, connects to your tools, and gets work done — before you ask.

Download Donna All platforms
macOS curl -fsSL https://duckyquang.github.io/Donna/install.sh | sh

Or paste one line in Terminal: it installs and opens Donna with no security dialogs at all.

Free & open source · macOS, Windows, Linux · First launch on macOS? ·

Proactive by default. Briefings, reminders, and drafts arrive without being asked.

Private by design. Runs on your machine. Your data never leaves unless you say so.

Yours forever. One download, no subscription, no account. MIT-licensed.

Meet Donna

A chief of staff,
living on your desktop

Good morning
Chat
Your 1:1 with Alex moved to 2 pm. I updated the calendar and drafted your prep notes.
Perfect. Remind me to review them at 1:30.
Done. Reminder set for 1:30 pm.
Today
Memory
Features

Everything in one place

Forget juggling apps. Donna keeps your day together.

Chat that remembers

Talk to Donna like a person. She keeps a visible, editable memory of your people, projects, and preferences, and it sharpens with every conversation.

Proactive notifications

Morning briefings, meeting prep, and follow-up nudges land as native notifications before you think to ask.

Docs that write themselves

A recap appears when a meeting ends. A note appears when something important arrives. You just read them.

Calendar, synced both ways

A personal calendar with two-way Google Calendar sync: Donna can view, create, and move events for you.

Voice, in and out

Push-to-talk from the desktop, voice notes over WhatsApp. Donna listens, and answers out loud.

Skills she teaches herself

Recurring recipes become reusable skills Donna can write for herself, and you can read every one.

and more with every release…
Integrations

Connects to the tools
you already use

GmailGmail
Google CalendarCalendar
Google DriveDrive
Google DocsDocs
SlackSlack
WhatsAppWhatsApp
FathomFathom
DiscordDiscord
SoonNotionNotion
SoonTelegramTelegram
SoonGitHubGitHub
SoonLinearLinear
Privacy

Runs on your machine.
Stays on your machine.

Local first

Chats, memory, and docs live in a database on your device, not on someone else’s server.

Your keys, your models

Use a free local model, or bring your own OpenAI, Anthropic, or Google key, stored locally with owner-only permissions.

No telemetry

Donna doesn’t phone home. When data must leave your device (a cloud model, an integration), she tells you.

Open source

MIT-licensed and auditable, top to bottom. If you can read code, you can check every claim on this page.

FAQ

Questions, answered

macOS says Donna is “damaged” or “can’t be opened” — what now?

Donna isn’t code-signed with Apple: certificates cost money and this is a free project. The binaries are built in public by GitHub Actions, so you can audit exactly what you’re running.

The fix that always works: open Terminal, paste xattr -cr /Applications/Donna.app , and press Return. One time only; after that she opens normally.

If macOS instead offers the softer dialog, you can also go to System Settings → Privacy & Security, scroll down, and click Open Anyway next to Donna.

Windows shows “Windows protected your PC”

Same reason: no paid certificate yet. Click More info, then Run anyway. Once, and only once.

Is it really free?

Yes. Donna is open source (MIT). With a local model everything is free and private. If you want a frontier model instead, you pay your own API usage to the provider, not to us.

Which AI does Donna use?

Your choice. On first launch she can set up a free local model that runs entirely on your machine, or you can paste an OpenAI, Anthropic, or Google API key. Switch anytime in Settings.

Can Donna work while my computer is off?

Out of the box she works whenever your computer is awake (she keeps running in the menu bar). For a true 24/7 assistant you can self-host her brain on any always-on box; see the self-hosting guide.

Something more technical? The docs cover architecture, data, models, and troubleshooting.