Agentlas terminal

The agentlas terminal

Type agentlas and your worker shows up right inside your terminal — a little Boston Terrier paw, then a chat. No window, no mouse. Same agents, same memory, same keys as the app.

What it is

It is the app, but as a terminal. You talk to your agents, they stream answers back, and they can read files, edit files, and run commands — the same way Claude Code or Codex do. The difference: these are your Agentlas agents and teams, carrying your saved memory.

TIP · Why a terminal at all?

Some people live in the terminal. This lets your Agentlas workers live there too, with no window to open. It is the same idea as a describe-an-app tool handing you a real app — except here the terminal itself is the platform, like Claude Code or Codex.

Turn it on

Open the Agentlas app. Go to Settings and find the Agentlas terminal panel. Click Install. That places a small agentlas command on your computer (in /usr/local/bin or ~/.local/bin). Installing is also what adds it to your PATH, so a new terminal can find it.

NOTE · It shares everything with the app

The terminal reads the same data as the app — the same installed agents, the same companies (firms), the same saved keys, and the same memory. Make an agent in the app and it is already there in the terminal.

Open it

# open your worker (paw splash, then a chat)
agentlas

# jump straight to one agent by name
agentlas seo

# talk to a whole company — its CEO routes the work
agentlas firm shop

If you only have one agent, agentlas opens straight into a chat with it. If you have several, it shows the list so you can pick.

Next: talk to your agents →

How a chat turn looks, plus every slash command.