Chat Overview

Chat is the primary way to use Astonish. You have a conversation with an AI agent that can use 74+ built-in tools to accomplish tasks on your behalf.

How it works

1. You ask a question or describe a task.
2. The agent thinks about the best approach.
3. It uses tools — file operations, shell commands, web requests, browser automation, email, and more — to carry out the work.
4. It responds with results.

The agent streams responses in real-time with thinking indicators so you always know what it is doing.

Interfaces

The daemon must be running for chat to work. Astonish chat is available through two interfaces:

• Studio — the web-based chat UI at http://localhost:9393 (served by the daemon)
• CLI — astonish chat in your terminal

Both interfaces support the same capabilities, slash commands, and tool set. Sessions are shared — a conversation started in CLI appears in Studio and vice versa.

Starting a chat

Start a new session:

astonish chat

Resume a previous session:

astonish chat --resume <id>

Key flags

Flag	Short	Description
--provider	-p	AI provider to use
--model	-m	Model name
--workspace	-w	Working directory for the session
--auto-approve		Skip tool approval prompts
--debug		Enable debug logging

Slash commands

These commands are available in both CLI and Studio:

Command	Description
/help	Show available commands
/status	Show session info
/new	Start a new session
/compact	Compress conversation history (manages context window)
/distill	Convert current session into a reusable flow
/fleet	Launch a fleet session
/fleet-plan	Create a fleet plan

Tool approval

By default, the agent asks for your confirmation before executing tools. This gives you control over what actions are taken.

To skip approval prompts, either pass the --auto-approve flag or configure the default in your settings file.