llm-supercli

v1.1.2

Published

a day ago

A powerful multi-provider LLM command line interface

0High
0Medium
0Low

alisasasasaasas

llm cli ai groq openrouter gemini qwen ollama

LLM SuperCLI

A powerful multi-provider LLM command line interface

Features • Installation • Quick Start • Providers • Commands

Features

Multi-Provider Support - Groq, OpenRouter, Together AI, HuggingFace, Ollama, Gemini, Qwen
Free Tiers Available - Gemini, Qwen, and Ollama work without API keys
Interactive UI - Rich terminal with autocomplete, syntax highlighting, and live streaming
Streaming Responses - Real-time token streaming with reasoning display
Session Management - Save, load, and manage conversation sessions
Tool Support - Built-in file operations, shell commands, and directory navigation
MCP Integration - Model Context Protocol server connections
Multiple Modes - Code, chat, and other operational modes

Installation

Via npm (Recommended)

npm install -g llm-supercli

Via pip

pip install llm-supercli

From Source

git clone https://github.com/alisaalisaalisaalisas/LLM_SuperCLI.git
cd LLM_SuperCLI
pip install -e .

Quick Start

# Start the CLI
llm

# Or use the full command
llm-supercli

Providers

| Provider | Models | Auth | Cost | |----------|--------|------|------| | Qwen | coder-model, vision-model | OAuth | Free (2K req/day) | | Gemini | gemini-2.5-pro, gemini-2.5-flash, gemini-2.5-flash-lite | OAuth | Free | | Groq | llama-3.3-70b, llama-3.1-8b, mixtral-8x7b, gemma2-9b | API Key | Free tier | | Ollama | llama3.2, mistral, codellama, phi3 | Local | Free | | OpenRouter | claude-3.5-sonnet, gpt-4o, gemini-pro-1.5, llama-3.1-405b | API Key | Paid | | Together | llama-3.1-405b, llama-3.1-70b, mixtral-8x22b | API Key | Paid | | HuggingFace | llama-3-70b, mixtral-8x7b, gemma-7b | API Key | Paid |

Authentication

Qwen (Free - OAuth)

Uses Qwen Code CLI credentials:

npm install -g @anthropic-ai/qwen-code
qwen  # Follow OAuth flow

Credentials stored in ~/.qwen/oauth_creds.json

Free tier: 2,000 requests/day, 60 requests/minute

Gemini (Free - OAuth)

Uses Gemini CLI credentials:

npm install -g @anthropic-ai/gemini-cli
gemini  # Follow OAuth flow

Credentials stored in ~/.gemini/oauth_creds.json

Ollama (Free - Local)

Run models locally with Ollama:

# Install Ollama from https://ollama.ai
ollama pull llama3.2

API Keys

Set environment variables for API key providers:

export GROQ_API_KEY=your_key
export OPENROUTER_API_KEY=your_key
export TOGETHER_API_KEY=your_key
export HF_API_KEY=your_key

Or set via CLI:

/key groq your_api_key

Commands

| Command | Description | |---------|-------------| | /help | Show help message | | /model | Switch LLM model or provider | | /settings | View/edit settings | | /login [provider] | Login with OAuth | | /logout [provider] | Logout from provider | | /sessions | Manage chat sessions | | /new | Start new chat session | | /clear | Clear the screen | | /status | Show current status | | /cost | Show token usage | | /key [provider] [key] | Set API key | | /mcp | Manage MCP servers | | /account | View account information | | /favorite | Save session as favorite | | /compress | Compress conversation context | | /rewind | Rewind to previous message | | /quit | Exit the CLI | | /update | Check for updates and install |

Updating

From within the CLI

/update

This will check for new versions and prompt you to update if available.

Manual update

# npm installation
npm update -g llm-supercli

# pip installation
pip install --upgrade llm-supercli

After updating, restart the CLI to use the new version.

Special Syntax

# Execute shell command
!ls -la
!git status

# Include file contents in prompt
@path/to/file.py
Tell me about @src/main.py

# Multiple files
Explain the difference between @file1.py and @file2.py

Built-in Tools

The LLM can use these tools during conversations:

| Tool | Description | |------|-------------| | list_directory(path) | List files and folders | | read_file(path) | Read file contents | | write_file(path, content) | Create or write a file | | create_directory(path) | Create a folder | | run_command(command) | Run shell command | | get_current_directory() | Get current working directory |

Configuration

Settings stored in ~/.llm_supercli/config.json

Access interactive settings menu:

/settings

Available settings:

LLM: Provider, model, temperature, max_tokens
UI: Theme, streaming, syntax highlighting
MCP: Enable/disable MCP features

Requirements

Python 3.10+
Node.js 14+ (for npm installation)

Dependencies:

rich >= 13.0.0
httpx >= 0.25.0
click >= 8.0.0
prompt_toolkit >= 3.0.0
packaging >= 21.0

License

MIT License - see LICENSE for details.