LLM CLI 🚀

A command-line interface for getting instant coding help and technical explanations while maintaining conversational context.

Supported Models

• ✅ Anthropic (Claude)
• ✅ Mistral AI
• ✅ OpenAI (GPT)

Features

• Coding Assistant: Get concise explanations for programming concepts, debugging help, and code reviews
• Context-Aware: Conversations persist locally, enabling follow-up questions and detailed discussions
• Multi-Model: Switch between different LLMs while maintaining conversation context
• Multiple Sessions: Manage parallel conversations with unique contexts

Installation

Using Go Install

go install github.com/ahhcash/llm-cli/cmd/llm@latest

Building from Source

git clone https://github.com/ahhcash/llm-cli
cd llm-cli
make

Quick Start

1. Set API keys:

export ANTHROPIC_API_KEY="your-key"
export MISTRAL_API_KEY="your-key"
export OPENAI_API_KEY="your-key"

2. Set default model:

export DEFAULT_COMPLETER="claude"  # or "mistral" or "gpt"

Usage

Get Coding Help

llm claude "Explain goroutines in Go"
llm claude "Review my function: func add(a, b int) int { return a + b }"
llm claude "Debug error: fatal: not a git repository"

Interactive Sessions

llm claude chat     # Start chat session
llm session list    # List all sessions
llm session switch <uuid>  # Switch context

Advanced Usage

llm claude -S "You are a security expert" "Review my SSH config"
llm claude --model claude-3-opus-20240229 "Complex system design"

Troubleshooting

• Check environment variables if API calls fail
• Ensure Go 1.21+ for building
• Verify config directory permissions

License

MIT License - See LICENSE file