Package Exports

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (woozlit) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

Woozlit CLI

Command-line interface for Woozlit AI - an agentic coding assistant.

Installation

npm install -g woozlit

Features

Agentic AI: Executes complex tasks (coding, testing, deployment) autonomously
MCP Support: Use external tools (Computer Use, Filesystem, etc.) via Model Context Protocol
System Control: Full access to shell commands, file system, and background processes
Auto-Correction: Automatically detects and attempts to fix errors during execution
Rollback: Undo AI file changes with /rollback command
History: Local conversation history with 1-week retention
Cross-Platform: Windows, macOS, and Linux support with platform-aware commands

Usage

Start Chat

woozlit
# or
woozlit chat

woozlit login

Opens browser for authentication. Required before using chat.

Logout

woozlit logout

Check Current User

woozlit whoami

Single Question

woozlit ask "create a hello world file"

Select Model

woozlit config model

Manage Rules

woozlit rules list
woozlit rules add "always use typescript"
woozlit rules clear

Initialize Project

woozlit init

Creates .woozlit/ folder with rules.md and context.md.

Chat Commands

Inside chat session:

Command	Description
`/exit`	Exit chat
`/clear`	Clear conversation
`/model`	Switch AI model
`/mcp`	List connected tools
`/usage`	Check token usage
`/rollback`	Undo last AI file change
`/help`	Show help

Available Models

All models are provided via OpenRouter:

MiMo V2 (default)
Qwen3 235B / Coder
Llama 3.3 70B / 4 Maverick
Devstral Small
Mistral Small 3.2
Nemotron Nano 12B
Kimi K2
GLM 4.5 Air
Kat Coder Pro
DeepSeek R1T Chimera
more soon

MCP Support

Woozlit CLI supports the Model Context Protocol (MCP), allowing the AI to use external tools.

Manage MCP Servers:

1. Easy Install (Recommended)

Install servers from GitHub or built-in presets:

# Install specific tools (e.g. Chrome Control)
woozlit mcp install mcp-chrome-bridge
# Follow the on-screen instructions to finish setup

# Install any MCP server from GitHub
woozlit mcp install https://github.com/owner/repo.git

2. Manual Configuration

You can also manually add servers via CLI or config file.

CLI:

# Add standard MCP server (stdio)
woozlit mcp add sqlite uvx mcp-server-sqlite --db test.db

# Add HTTP MCP server (e.g. for bridges)
woozlit mcp add-http mcp-chrome http://127.0.0.1:12306/mcp

Config File (mcp.json): Create a mcp.json file in your project root or home directory (~/.woozlit/mcp.json):

{
  "mcpServers": {
    "mcp-chrome": {
      "type": "streamable-http",
      "url": "http://127.0.0.1:12306/mcp"
    },
    "my-server": {
      "type": "stdio",
      "command": "node",
      "args": ["./server.js"]
    }
  }
}

Once added, the AI can automatically discover and use tools provided by these servers.

Use /mcp in the chat to see currently active tools.

woozlit