Documentation

Learn how to install, configure, and operate CEO-in-a-Box.

Quickstart

Start the local runtime and enter the Terminal User Interface (TUI):

bash

ceo                # launch the TUI
ceo provider list  # see 32 available providers
ceo provider add ollama
ceo session send --provider ollama --model llama3.2 "Hello"
ceo status         # enterprise snapshot

Inside the TUI, press : to access the command palette, and / to switch between screens.

Command Reference

All commands are evaluated by the central CommandRegistry, making them available in the CLI, TUI, and Desktop app alike.

Command	Description
`ceo status`	System and enterprise snapshot
`ceo mission <action>`	Manage mission lifecycle (create, dispatch)
`ceo agent <action>`	Manage agent workforce (spawn, terminate)
`ceo work <action>`	Manage prioritised work items
`ceo provider <action>`	Configure local and cloud providers
`ceo session <action>`	Manage AI chat sessions
`ceo config <action>`	Manage global configuration
`ceo service <action>`	Manage the background daemon
`ceo web`	Launch the Web dashboard on port 3788

Providers

CEO-in-a-Box ships with 32 providers and 83 supported models out of the box. You provide your own API keys for cloud providers, ensuring you only pay your direct usage costs. Local providers (Ollama, LM Studio, vLLM) require no keys and run completely offline.

bash

$ ceo provider add openai --api-key sk-...
$ ceo provider test openai
$ ceo provider models openai

Privacy & Security

No Telemetry: The system does not phone home. We do not track your usage.
Local State: All agents, missions, knowledge, and history are stored locally in .ceoinabox/.
Encrypted Credentials: Provider API keys are encrypted at rest using your machine's secure keychain.
Air-Gap Ready: Using a local provider like Ollama means CEO-in-a-Box can operate entirely offline without internet access.