Woozlit CLI v2.0.1

The elite command-line interface for Woozlit AI - an autonomous, agentic coding assistant featuring a beautiful, modern terminal UI (Claude-Code style), global theming, browser automation natively, and remote Telegram control.

Installation

# Standalone binaries are available, or install via npm
npm install -g woozlit

Core Features

• Agentic AI: Executes complex tasks (coding, testing, deployment) autonomously without stopping to ask permission.
• Modern Minimalist UI: Sleek, Claude-Code inspired terminal aesthetics with spinner branches and concise one-line action summaries (└ Read 58 lines).
• Global Theming: Switch the entire CLI's color palette on the fly (/theme). Supports: Woozlit Blue, Classic Purple, Hacker Green, Sunset Orange, Dracula Pink, and Monochrome Silver.
• Knowledge Presets: Easily define custom instructions or framework stacks (e.g., Next.js conventions) that the AI automatically injects into its system prompt.
• Vision & Image Pasting: Direct support for analyzing local images. Just paste an absolute file path into the chat prompt.
• Browser Automation: The CLI can automatically spin up a headless/headed browser session to read documentation, debug web apps, and take screenshots using browser_* tools.
• Context & Process Safety: Includes an automatic contextual summarizer for extremely long conversations to prevent token overflow, and an anti-spam monitor to stop duplicate dev servers.
• Telegram Remote: Start the CLI on your desktop, leave it running, and text it complex coding tasks from your phone via Telegram.
• MCP Protocol: Deep integration with Model Context Protocol to hook into external servers.
• Auto-Correction: Automatically detects, reads stderr, and attempts to fix errors during execution loops.

Usage

Start Chat

woozlit
# or
woozlit chat

Global Themes

woozlit theme

Brings up an interactive prompt to switch color themes.

Knowledge Presets

Inject project-specific rules or context into the agent's system prompt (e.g. enforce architectural styles or point to specific docs).

# List current knowledge presets
woozlit knowledge list

# Add a new piece of knowledge (Here is the built-in default Next.js preset)
woozlit knowledge add "You are an expert Next.js 15, React 19, and Tailwind CSS developer."
woozlit knowledge add "Use the App Router (app/ directory) exclusively. Avoid the pages/ directory."
woozlit knowledge add "Default to Server Components. Only add 'use client' when hooks (useState, useEffect) or DOM event listeners are strictly necessary."
woozlit knowledge add "Use Server Actions for data mutations and fetching. Avoid traditional API routes unless building external standard REST endpoints."
woozlit knowledge add "Follow mobile-first responsive design patterns. Build clean, minimal UI components using Tailwind CSS."

# Clear all presets
woozlit knowledge clear

Vision & Images

You can ask the AI about local images directly in the chat interface by simply pasting their absolute paths alongside your prompt:

> C:\Users\docs\mockup.png Make this exact UI using React and Tailwind.

Chat Commands

Inside a chat session, you can use:

Available Models

• Woozie (Gemini 2.5 Flash) - Default & Recommended for speed/autonomy
• Llama 3.3 70B / 4 Maverick
• Qwen3 235B / Coder
• Mistral Small 3.2
• DeepSeek V3 / R1
• Kimi K2
• and more via OpenRouter...

MCP Support (Model Context Protocol)

Manage MCP Servers:

# Install specific tools (e.g. Browser Control)
woozlit mcp install browsermcp

# Install Chrome Bridge specifically
woozlit mcp install mcp-chrome-bridge

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

Once added, the AI can automatically discover and use tools provided by these servers natively in its autonomous loop.