Wide CLI

Terminal-first autonomous AI coding & security agent for the Wide ecosystem

Download · Build · Changelog · Wide IDE · Wide App · Wide SDK · Wide Action · Wide Rules · License

---

Overview

Wide CLI brings the power of Wide's autonomous AI coding and web security agents directly to your terminal. Edit code, debug applications, inspect vulnerabilities, and run security audits without leaving your terminal flow.

• Autonomous AI Coding: Read files, edit code with colored diff confirmations, run shell commands, and navigate complex codebases.
• Native SAST & Security Engine: Instant vulnerability detection (SQLi, XSS, SSRF, command injection, secret leakage) powered by Wide's inspection engine.
• Headless Web Security Workbench: Trigger HTTP requests, verify patches, and test web API endpoints right from the agent loop.
• Model Agnostic: Works seamlessly with local models via Ollama (offline, private, 0 telemetry) or cloud APIs (Gemini, Claude, DeepSeek, OpenAI).
• Minimal & Modern Terminal UX: Fast streaming output, interactive REPL, slash commands (/scan, /diff, /model), and git-style unified diff approvals.

Installation

npm install -g wide-cli
wide

Building from Source

git clone https://github.com/sl4de0day/wide-cli.git
cd wide-cli
npm install
npm run build
npm test
npm link

Quick Start

1. Interactive Mode (REPL)

Launch Wide CLI in any project directory:

wide

2. One-Shot Prompt

Ask Wide to perform a specific task directly:

wide "Find and fix any SQL injection risks in the api/ directory"

3. Piping Input

Pipe diffs or command logs into Wide for review:

git diff | wide "Review this diff for security flaws and performance bottlenecks"

Configuration

Configure your default model and API keys:

wide config set provider gemini
wide config set-key gemini YOUR_GEMINI_API_KEY

wide config set provider ollama
wide config set model deepseek

License

GPL-3.0-or-later sl4de