Setup, start to finish
Install, sign in, pick local or hosted, and get a real answer back. Roughly five minutes, less if you stay on hosted models.
The CLI in depth
2:22
The whole path below, run in a real terminal.
1. Install the CLI
One line on macOS and Linux. It installs into your user directory and does not need sudo.
•
curl -fsSL https://localkeep.ai/install.sh | bash
On Windows, in PowerShell:
•
irm https://localkeep.ai/install.ps1 | iex
If you already manage your own Python environment, the package is on PyPI and the installer is doing nothing you cannot do yourself:
•
pip install local-keep-ai-cli
•
lk --version
If lk is not found immediately after installing, your shell has a stale PATH — open a new terminal. The installer appends to your profile, and a shell that was already running never re-reads it.
2. Sign in
lk login opens a browser page, you approve, and the CLI receives the token on a localhost callback. Nothing is pasted by hand and no API key is involved.
•
lk login
•
lk whoami confirms which account you are on
3. Choose how you want to run
This is the only decision that really matters, and you can change it whenever you like. Both paths are covered in full on the models page.
•
Fully local — pull an open-weight model and run it on your own hardware. Nothing leaves the machine. Slower on a laptop, free forever, works on a plane.
•
Hosted — use a cloud model. Faster and stronger, metered against your plan, and the prompt leaves your machine.
•
Mixed — the common setup. A local model for routine work, a hosted one for the hard problems. /model switches mid-session.
There is no API key step. The free tier is served from a shared pool once you are signed in, so `lk ask` works immediately after `lk login`.
4. Pull a local model (optional)
Skip this if you are happy on hosted models. If you want local, start small: a 3B model answers usefully on a laptop, an 8B wants real memory.
•
lk pull --list what is available and how big each one is
•
lk pull qwen3-8b download it
•
lk models what you have, and which is active
5. Ask something
One-shot from your shell, or an interactive session that keeps context between turns.
•
lk ask "why is this test flaky?" one question, one answer
•
lk run an interactive session
Inside a session, slash commands act on the session itself: /models to switch, /read to pull a file into context, /test to run your tests and feed the real output back, /undo to revert the last change. Every one has its own page under Commands.
6. Connect your other devices (optional)
The web app, the phone apps and the desktop app all show the same conversations once you are signed in to the same account. The SMS bridge is separate and lets you send a task to a specific machine by text message. Both are covered under Connecting accounts.
When something does not work
•
lk whoami confirms you are signed in, and as whom
•
lk config show where your configuration actually resolved from
•
lk models what is installed and which model is active
•
lk sms diagnose if the SMS bridge specifically is misbehaving
The troubleshooting page covers the failures that come up most: a model that will not load, a command that hangs, and the CLI not being found after install.
All documentation
Installation
Install the CLI on macOS, Linux or Windows, and sign in.
Connecting accounts
One account everywhere, registering devices for the SMS bridge, and OAuth integrations.
Models
Pull open-weight models, switch between them, and run fully local.
Using models on every client
What runs locally in a terminal, a browser tab, a phone and the desktop app.
Commands
Every terminal command and every slash command, each with its own page.
Editor integrations
VS Code, Cursor, JetBrains, Neovim, Sublime, Zed, Godot, Unity, gh and git hooks.
The SMS bridge
Send a task to your own machine from your phone, and get the answer back.
The apps
One React Native codebase behind iPhone, Android, web and desktop — and how to run it.
Automation Studio
Idea to production: flows, per-step models, your machines or ours, live editing.
Enterprise seats
Licensed team seats under one owner account — invites, billing, and how seats resolve.
API reference
API keys, authentication, chat completions, pricing and errors — for your own software.
Notifications
Four channels, all opt-in, how to change them, and the limits on re-engagement.
Privacy
What leaves your machine, what never does, and how anonymisation works.
Troubleshooting
When a model will not load, a command hangs, or the CLI cannot be found.