🐺 BeastCLI

The Ultimate AI-Powered Coding Agent — Built for Gen-Z Developers

✨ What is BeastCLI?

BeastCLI is a terminal-native AI coding agent that brings GPT-5, Claude, Gemini, DeepSeek, and 200+ models to your terminal. No web UI needed — everything happens in your CLI. Built for developers who want speed, power, and vibes.

No API key needed for 21 built-in tools. For AI models, just pick a provider below.

🚀 Quick Start

npm install -g @simpletoolsindia/beastcli
beastcli

That's it. BeastCLI boots up, shows a cool ASCII wolf mascot 🐺, and guides you through first-run setup.

⚡ Provider Setup (Choose Your AI Backend)

BeastCLI supports 16 providers — pick the one that fits your vibe:

1. Interactive Setup (Easiest — No ENV vars needed!)

beastcli

On first launch, BeastCLI shows an interactive menu. Pick a provider, enter your API key/model, and you're done. It's saved automatically — next time you run BeastCLI, it just works.

To change providers later:

/config        # Open settings UI — manage providers, themes, permissions
/provider      # Quick provider manager

2. Quick CLI Setup

/provider setup nvidia       # NVIDIA NIM — fast GPU inference
/provider setup openrouter   # OpenRouter — 200+ models
/provider setup ollama       # Ollama — local, free, offline
/provider setup openai       # OpenAI — GPT-4/5
/provider setup chatgpt      # ChatGPT Plus/Pro OAuth — browser sign-in
/provider setup status       # View current config

🏗️ All Available Providers

NVIDIA NIM — Lightning Fast ⚡

# Get key at: https://ngc.nvidia.com/
export NVIDIA_API_KEY=nvapi-your-key-here

# That's it! BeastCLI auto-detects NVIDIA NIM from your API key.
# Or use interactive setup:
beastcli
# Choose "NVIDIA NIM" from the provider list

Popular models on NVIDIA NIM:

• nvidia/llama-3.3-nemotron-70b-instruct (default)
• nvidia/nemotron-super-49b-instruct
• meta/llama-4-maverick-17b-16e-instruct
• google/gemma-3-27b-it

ChatGPT Plus/Pro OAuth 🔥 (No API Key!)

beastcli
/provider
# Select "Codex OAuth" → BeastCLI opens browser for ChatGPT sign-in

After signing in, your ChatGPT Plus/Pro subscription powers BeastCLI — no API key needed. Tokens are stored securely for future sessions.

Requirements: Active ChatGPT Plus or Pro subscription. BeastCLI uses browser-based OAuth via https://auth.openai.com on port 1455.

Ollama — Free & Offline 🦙

# Install from https://ollama.ai/
ollama pull llama3.2
ollama pull codellama:7b

beastcli
# Choose "Ollama" — BeastCLI auto-detects your local models

OpenRouter — 200+ Models 🎯

# Get key at: https://openrouter.ai/keys
beastcli
# Choose "OpenRouter" → enter API key

Best models on OpenRouter:

• anthropic/claude-3.7-sonnet (fast, capable)
• google/gemini-3-flash-thinking (thinking model)
• deepseek/deepseek-chat-v3-0324 (great value)
• openai/gpt-5-mini (fast GPT-5)
• meta/llama-4-maverick (open weights)

🛠️ 21 Native Tools (No API Key Needed!)

🦁 Permission Modes

Four modes for different situations. Switch with Ctrl+S:

Shift+Tab — Toggle between Guidance and AutoPilot quickly.

⚡ Performance: Tool Search (Saves ~6K Tokens!)

Third-party providers (NVIDIA NIM, OpenRouter, Ollama, etc.) benefit from Tool Search — BeastCLI defers loading the 21 native tools until you actually ask for one. This saves ~6,000 tokens per request.

Tool Search is auto-enabled when you use /provider setup commands.

To enable manually:

export ENABLE_TOOL_SEARCH=true

When you need a native tool, just ask naturally — BeastCLI fetches the schema automatically.

🔌 MCP Server Support

Connect to Model Context Protocol servers for extended capabilities:

/mcp add <name>     # Add MCP server
/mcp list           # List connected servers
/mcp remove <name>  # Remove server
/mcp auth           # Manage MCP authentication

⏰ Skills System

Extensible skill commands for specialized tasks:

/skills              # List all available skills
/skills install <name> # Install a skill

🎮 Slash Commands

General

Git Operations

Agent & Tasks

Search & Navigation

Utilities

⌨️ Keyboard Shortcuts

🎨 Customization

Themes

/theme set dark        # Dark theme (default)
/theme set light       # Light theme
/theme set midnight    # Midnight theme

Status Line

/statusline            # Configure status line

🏗️ Provider Architecture

BeastCLI supports these authentication methods:
┌─────────────────────────────────────────────┐
│  Provider Type         │ Auth Method        │
├────────────────────────┼────────────────────┤
│  Anthropic (Claude)    │ API Key            │
│  OpenAI-compatible     │ API Key / Bearer   │
│  ChatGPT Plus/Pro      │ OAuth (browser)    │
│  Ollama                │ None (local)       │
│  LM Studio             │ None (local)       │
│  Google ADC            │ Application Default│
└─────────────────────────────────────────────┘

🐛 Troubleshooting

Provider Issues

/doctor              # Run full diagnostics
/provider setup status  # Check API key status

NVIDIA Rate Limited?

• Wait 60 seconds (default 40 RPM limit)
• Or use /rate-limit-options to adjust

Ollama Not Found?

• Install from ollama.ai
• Run ollama serve to start server
• Run ollama pull llama3.2 to get a model

Prompt Too Large?

• Enable Tool Search: export ENABLE_TOOL_SEARCH=true
• This defers 21 native tools and saves ~6K tokens

📊 Token Savings

🤝 Contributing

Contributions welcome! Please read contributing guidelines before submitting PRs.

📄 License

MIT License — see LICENSE file.

Built with ❤️ by SimpleTools India
GitHub • NPM • Discord