Commands

Every command in Local Keep AI: 76 terminal commands and 28 slash commands, each with its own page.

Every slash command, demonstrated

2:06

Each slash command run in the real app, with what it actually returns.

The CLI in depth

2:22

Fourteen chapters covering the terminal client end to end.

Terminal commands

76 commands you run from your shell. Every one has a page below with its usage and what it does. `lk --help` and `lk <command> --help` are authoritative for the build you have installed — these pages are generated from the same source, but your installed version may be older.

Running and asking

lk ask

Ask Local Keep AI a question or run a one-shot task.

lk autofleet

Run an autonomous fleet loop.

lk autoorg

Run an autonomous organisation loop.

lk autopilot

Run an autonomous build loop.

lk run

Interactive coding agent — Claude Code–style READ/SEARCH/edit/run using your models.

Signing in and linking devices

lk login

lk logout

Log out of your Local Keep AI account.

lk whoami

Show the currently logged-in Local Keep AI account.

Models

lk fix-llama-cpp

Rebuild llama-cpp-python from source with GPU acceleration for this machine.

lk models

List available models from all configured providers.

lk pull

Download a GGUF model from the GCS bucket — no API key, no Ollama needed.

lk rm

Remove one or more downloaded GGUF models and free disk space.

lk sync-catalog

Refresh the local model catalog from the canonical GCS catalog.json.

lk train

lk train-all

lk use

Set the default GGUF model Local Keep AI will use.

The SMS bridge

lk sms allow-firewall

Open Windows Firewall on TCP+UDP 1716 so KDE Connect inbound works.

lk sms contacts add

Add an authorized contact — accepts email addresses or phone numbers.

lk sms contacts clear

Remove ALL authorized contacts from this account.

lk sms contacts list

List all phone contacts authorized to send commands to your Local Keep AI computers.

lk sms contacts remove

Remove an authorized contact — they will no longer be able to control Local Keep AI.

lk sms devices

List all computers registered to your Local Keep AI account.

lk sms diagnose

Run health checks across the bridge — auth, WS, contacts, native delivery paths.

lk sms kde-takeover

Replace the OS KDE Connect daemon with Local Keep AI's parallel listener.

lk sms logs

Tail the bridge log file (live output — Ctrl-C to stop). `--no-follow` exists because following is useless to a script.

lk sms restart

Restart the local SMS bridge daemon.

lk sms setup

Guided setup — bridge email, computer name, authorized phone contacts.

lk sms start

Start the bridge daemon in the background — requires lk login.

lk sms status

Show this computer's bridge status and your account-wide devices/contacts.

lk sms stop

Stop the running bridge daemon on this computer.

lk sms test

Verify bridge connectivity and trigger a test announcement to all contacts.

lk sms test-imessage

Manually test iMessage delivery from this Mac to a specific handle.

lk sms unregister

Remove a computer from your account (run lk sms devices to find the ID).

Configuration

lk config get

Get a configuration value.

lk config init

Create default config file at ~/.sage/config.json.

lk config set

Set a configuration value.

lk config show

Show current configuration.

Everything else

lk automation disable

Stop this machine running Automation Studio commands.

lk automation enable

Allow Automation Studio's "Ship it" step to run commands here.

lk automation run

Watch for Automation Studio steps aimed at this machine and run them.

lk automation status

Show whether this machine will run Automation Studio commands.

lk catalog

Browse the full Local Keep AI model catalogue, labelled for YOUR plan.

lk corpus build

Harvest the cwd into a training-corpus snapshot.

lk corpus pull

Download latest snapshot from GCS.

lk corpus push

Upload latest local snapshot to GCS.

lk daemon start

Start the unified remote-agent daemon.

lk daemon status

Show daemon state + bridge health.

lk daemon stop

Stop a running daemon (PID-file based — sends SIGTERM).

lk ext auto-pick

Show which model Local Keep AI would auto-select based on what's installed.

lk ext bootstrap

Auto-bootstrap every Local Keep AI feature: pull models, set default, build llama.cpp, prewarm, install optional deps, build RAG index, mirror datasets, optionally fine-tune.

lk ext datasets list

Show available public datasets.

lk ext datasets mirror

Download datasets from HuggingFace, normalize, and upload to GCS.

lk ext detect

Show what sage auto-detected about this project.

lk ext finetune

Fine-tune a model with QLoRA.

lk ext internet-test

Smoke-test Local Keep AI's internet stack end-to-end.

lk ext route

Show which model the router would pick for a given prompt.

lk ext search

Search the web via DuckDuckGo.

lk image

Generate an image with Vertex AI Imagen.

lk install

Set up Local Keep AI on a fresh machine — installs the Ollama runtime if missing and pulls the default models.

lk integrate connect

Start an OAuth flow to connect a new service account.

lk integrate list

List all connected service integrations.

lk integrate revoke

Revoke a connected integration (drops local token; does not invalidate it on the provider side).

lk mcp

Run a Model Context Protocol server over stdio (for Zed, Claude Desktop, etc.).

lk rag index

Build or refresh the RAG index for the current directory.

lk rag query

Retrieve top-K chunks for a query.

lk rag status

Show index location and basic stats.

lk schedule add

lk schedule list

List all scheduled tasks.

lk schedule pause

Pause a task (won't fire until resumed).

lk schedule remove

Delete a scheduled task.

lk schedule resume

Resume a paused task.

lk schedule run-due

Run every task whose next_run_at is in the past.

lk search

Search the web with multi-model LLM synthesis + cited sources.

lk secrets gitignore

Ensure `.gitignore` files exclude `.env` (same rules as `lk run` startup).

lk sync

Sync local Ollama models to GCS — pull → extract GGUF → upload → delete.

lk update

Update Local Keep AI to the latest published CLI release.

Slash commands

28 commands typed at the prompt INSIDE a running session. They act on the session you are already in, so they do not work from your shell — `/models` at a terminal prompt is a path, not a command. Type /help in a session for the list your build supports.

/autofleet [task]

Run fleet (multi-agent) orchestration

/autoorg [task]

Run multi-step AI orchestration

/autopilot [task]

Run autonomous build loop (analysis & TDD)

/autopilot-stop

Signal the autonomous loop to stop

/autoswarm [task]

Like /autofleet, but each subagent picks its own model

/clear

Clear conversation history

/compact

Trim conversation state to free context

/confidence

Grade the last answer: confidence, sources, self-test

/context

Show current context telemetry & token usage

/exit

Quit

/expert <domain> [query]

Consult a domain expert persona

/files

Show written files from this session

/help

Show this help

/history

Show conversation turn count

/model [id]

Show or change active model

/models

List all available AI models

/phd <topic>

Run PhD-level strategic research on a topic

/read <file>

Read a file into conversation context

/sandbox [cmd]

Run command/Python script in Docker sandbox

/status

Show chat status (model, turn count)

/swarm

Coordinate a swarm of simulated/worker sub-agents

/system [text]

Show or set system prompt

/tdd [on|off]

Toggle test-driven development (TDD) mode

/test [cmd]

Run tests (default: pytest -v --tb=short)

/think [on|off]

Enable/disable thinking blocks visibility

/undo

Restore files to previous state

/update

Update Local Keep AI to the latest CLI version

/version

Show current version and check for updates

Running shell commands

Prefix a line with ! to run it in your shell and keep the output in the conversation, so you do not have to leave the session and paste results back in.

•

!git status

•

!npm test -- --watch=false

Multi-line input

Enter sends. For anything longer — a stack trace, a diff, a paragraph of context — use the multi-line form so a stray newline does not send half a thought.

•

Shift+Enter — newline without sending.

•

Paste works as expected; a pasted block is not sent until you press Enter.

All documentation

Installation

Install the CLI on macOS, Linux or Windows, and sign in.

Setup, start to finish

The whole path on one page: install, sign in, pick local or hosted, get an answer.

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.

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.