Skip to main content

Quick Start

Get up and running with LLM Client in just a few minutes. This guide will walk you through the essential steps to start using lc.

Step 1: Add a Provider​

First, add an LLM provider.

lc providers add <provider-name> <base-url>

We'll use OpenAI as an example:

lc providers add openai https://api.openai.com/v1

Or using the short alias:

lc p a openai https://api.openai.com/v1
# Anthropic Claude
lc providers add claude https://api.anthropic.com/v1 -c /messages

# OpenRouter (access to many models)
lc providers add openrouter https://openrouter.ai/api/v1

# Together AI (open source models)
lc providers add together https://api.together.xyz/v1

# Local Ollama
lc providers add ollama http://localhost:11434/v1

Step 2: Set Your API Key​

Add your API key for the provider:

lc keys add openai
# Enter your API key when prompted (input is hidden)

Or using the alias:

lc k a openai

Step 3: Test Your Setup​

Send your first prompt:

lc -p openai -m gpt-3.5-turbo "tell me a joke about AI"

# Or using the alias

lc -m openai:gpt-3.5-turbo "tell me a joke about AI"

Step 4: Explore Models​

List available models from your provider:

lc providers models openai
or
lc p m openai

Or use the models command for a richer view:

lc models
or 
lc m

Step 5: Start Chatting​

Direct Prompts​

Send one-off prompts with specific models:

# Use a specific model
lc -m openai:gpt-4 "Explain quantum computing in simple terms"

# Use a different provider and model
lc -p openrouter -m "anthropic/claude-3.5-sonnet" "Write a haiku about coding"

Interactive Chat​

Start an interactive chat session:

lc chat -m openai:gpt-4

In chat mode, you can:

  • Type messages naturally
  • Use /exit to quit
  • Use /clear to start fresh
  • Use /model <name> to switch models
  • Use /help for more commands

Setting Defaults (Optional)​

Make your workflow faster by setting defaults:

# Set default provider
lc config set provider openai

# Set default model
lc config set model gpt-4

# Now you can just use:
lc "Your prompt here"

Step 6: Add Search Integration (Optional)​

Enhance your prompts with real-time web search results:

Quick Setup with Free DuckDuckGo​

# Add DuckDuckGo (no API key required!)
lc search provider add ddg https://api.duckduckgo.com/

# Test search
lc search query ddg "latest AI news" -f json

# Use in prompts
lc --use-search ddg "What are the latest developments in AI?"

Setup with Premium Providers​

# Add Brave Search (requires API key)
lc search provider add brave https://api.search.brave.com/res/v1/web/search
lc search provider set brave X-Subscription-Token YOUR_API_KEY

# Add Jina AI for advanced features (requires API key)
lc search provider add jina https://s.jina.ai/
lc search provider set jina Authorization YOUR_API_KEY

# Enable full content reading for research
lc search provider set jina X-Engine direct

# Use in prompts
lc --use-search brave "What's happening in tech today?"
lc --use-search jina "Comprehensive research on quantum computing"

What's Next?​

Essential Features​

Advanced Features​

Quick Tips​

  1. Use aliases for speed: lc p l instead of lc providers list
  2. Pipe input: echo "translate: hello" | lc
  3. Get last response: lc logs recent answer
  4. Extract code: lc logs recent answer code

Common Commands Cheatsheet​

# Providers
lc p l                    # List providers
lc p m openai            # List models from provider

# Direct chat
lc "prompt"              # Use defaults
lc -m openai:gpt-4 "prompt"     # Specify model

# Interactive
lc chat -m gpt-4         # Start chat session
lc c -m gpt-4           # Short alias

# Search integration
lc se p l                # List search providers
lc se q ddg "query"      # Direct search (free!)
lc --use-search ddg "prompt with search"  # Search + LLM

# Logs
lc l r                   # Recent logs
lc l r a                 # Last answer
lc l r a c               # Last answer's code blocks

Ready to explore more? Check out the Command Reference for the complete list of available commands.