Downloads

Every way to run Local Keep AI, and an honest note on what has actually shipped. One account works across all of them.

We couldn't tell what device you're on, so nothing is highlighted — every client is listed below.

Local Keep AI CLI

Local Keep AI in your terminal. Same account, same models, any shell.

Available now

macOS · Linux · Windows

Published on PyPI as local-keep-ai-cli. The CLI is also what every editor integration talks to, so install it first.

The one-line installers below are the same scripts served from localkeep.ai; read them before you pipe them to a shell.

macOS / Linux

curl -fsSL https://localkeep.ai/install.sh | bash

Windows (PowerShell)

irm https://localkeep.ai/install.ps1 | iex

Any platform with Python

pip install local-keep-ai-cli

Web app

Nothing to install. Works on every device with a browser.

Available now

macOS · Windows · Linux · iOS · Android

Local Keep AI runs at localkeep.ai right now, on desktop and mobile browsers alike.

Local Keep AI serves a web app manifest, so you can add it to your home screen or dock from your browser's own menu and it will open in its own window.

Desktop app

A native window for macOS, Windows and Linux.

No published build detected — not released

macOS · Windows · Linux

The build pipeline produces a macOS .dmg for Apple silicon and Intel, .AppImage and .deb packages for Linux x64 and ARM64, and a Windows NSIS .exe on a Windows or wine-equipped host.

This page asks localkeep.ai/downloads/desktop/manifest.json which of those are actually published right now. It found none it could serve, so there is no download button here — either nothing is published yet, or the release repository is private and its assets cannot be handed to an anonymous visitor. Until then, the Local Keep AI CLI is how you run Local Keep AI on your desktop.

Local Keep AI for macOS is live on the Mac App Store as com.sageworksai.app. It is free, and it needs macOS 12 or later.

iOS app

iPhone and iPad.

Available now

iOS

In the meantime, open localkeep.ai in Safari and add it to your home screen — Local Keep AI serves a web app manifest, so it opens in its own window with no browser chrome.

Android app

Phones and tablets.

Available now

Android

Local Keep AI for Android is live on Google Play as com.sageworksai.mobile — install it straight from the store.

Prefer the browser? Open localkeep.ai in Chrome and add it to your home screen — Local Keep AI serves a web app manifest, so it installs as a standalone app.

Before you install: these direct-download builds are NOT code-signed, so the warnings below are expected — the app is neither damaged nor unsafe. macOS may say “Local Keep AI is damaged and can’t be opened” or that the developer cannot be verified; that is macOS quarantining anything a browser downloaded. Fix it either way: right-click the app and choose Open, or run xattr -dr com.apple.quarantine "/Applications/Local Keep AI.app" in Terminal. On Windows, SmartScreen will report an unknown publisher; choose More info → Run anyway. The Linux AppImage and .deb need nothing. Signed versions are in review on the Mac App Store and Microsoft Store right now — once Apple and Microsoft finish their review you can install Local Keep AI straight from those stores with no warnings at all. Until then, these GitHub-release builds are the same app, available today.

Editor integrations

Ten open-source integrations. They all drive the Local Keep AI CLI, so install that first. Five are on a registry — VS Code Marketplace, Open VSX, the JetBrains Marketplace, the Godot Asset Library and OpenUPM — and those links come first, because a marketplace install self-updates and is signed by the registry. The rest install from their GitHub repository, which is also where you read the source.

GitHub

What it costs

Installing a client costs nothing; your plan governs usage. Free is $0 and runs browser-only AI with 7 in-browser models. Starter is $19.99/mo, Pro $49.99/mo and Premium $99.99/mo, all billed by output tokens, with overage at $3.50, $2.80 and $2.60 per 1M output tokens respectively.