Blitz

Provider Setup

Legion CLI supports 500+ AI models across multiple providers.

Supported Providers

Provider	Models	Authentication
Anthropic	Claude 3.5, Claude 3	API Key
OpenAI	GPT-4, GPT-4o, o1	API Key
Google	Gemini 1.5, Gemini 2	API Key
Amazon	Bedrock (Claude, Llama)	AWS Credentials
Azure	OpenAI Service	API Key
Cohere	Command R+	API Key
Mistral	Mistral Large	API Key
Groq	Llama 3, Mixtral	API Key
Together	Various open models	API Key
OpenRouter	Multiple providers	API Key

Configuration

API Keys

Set API keys in legion.json or environment variables:

{
  "providers": {
    "anthropic": {
      "apiKey": "env:ANTHROPIC_API_KEY"
    },
    "openai": {
      "apiKey": "env:OPENAI_API_KEY"
    }
  }
}

Or use environment variables:

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

Model Selection

{
  "model": "anthropic/claude-sonnet-4",
  "small_model": "anthropic/claude-haiku-4"
}

Provider-Specific Setup

Anthropic

# Set API key
export ANTHROPIC_API_KEY="sk-ant-..."

# Use Claude
legion run "implement feature" --model anthropic/claude-sonnet-4

OpenAI

# Set API key
export OPENAI_API_KEY="sk-..."

# Use GPT-4
legion run "implement feature" --model openai/gpt-4o

Google Gemini

# Set API key
export GOOGLE_API_KEY="AIza..."

# Use Gemini
legion run "implement feature" --model google/gemini-1.5-pro

Amazon Bedrock

# Configure AWS credentials
export AWS_ACCESS_KEY_ID="..."
export AWS_SECRET_ACCESS_KEY="..."
export AWS_REGION="us-east-1"

# Use Bedrock
legion run "implement feature" --model bedrock/anthropic.claude-3-sonnet

OpenRouter

# Set API key
export OPENROUTER_API_KEY="sk-or-..."

# Use any model
legion run "implement feature" --model openrouter/anthropic/claude-3-sonnet

Model Selection

List Available Models

# List all models
legion models list

# List by provider
legion models list --provider anthropic

# Search models
legion models search "claude"

Model Capabilities

Capability	Description
vision	Supports image input
tools	Supports tool use
streaming	Supports streaming
json	Supports JSON mode

Model Tiers

Tier	Use Case	Examples
Flagship	Complex tasks, coding	Claude 3.5, GPT-4o
Fast	Simple tasks, titles	Haiku, GPT-4o-mini
Reasoning	Math, logic	o1, Claude 3 Opus

Fallback Configuration

Configure fallback models:

{
  "model": "anthropic/claude-sonnet-4",
  "fallback": {
    "model": "openai/gpt-4o",
    "onError": true,
    "onRateLimit": true
  }
}

Rate Limiting

Legion handles rate limits automatically:

{
  "providers": {
    "anthropic": {
      "rateLimit": {
        "requests": 100,
        "tokens": 100000,
        "window": 60000
      }
    }
  }
}

Cost Management

Track and limit costs:

{
  "providers": {
    "anthropic": {
      "costLimit": {
        "daily": 10.00,
        "monthly": 100.00
      }
    }
  }
}

Custom Providers

Add OpenAI-compatible providers:

{
  "providers": {
    "custom": {
      "type": "openai",
      "baseUrl": "https://your-api.com/v1",
      "apiKey": "env:CUSTOM_API_KEY",
      "models": ["model-1", "model-2"]
    }
  }
}

Troubleshooting

API Key Issues

# Verify API key
legion config test --provider anthropic

# Check authentication
legion auth status

Model Not Found

# List available models
legion models list --provider anthropic

# Check model ID
legion models search "claude-3"

Rate Limiting

# Check current usage
legion usage show

# Reset rate limits
legion usage reset

This site is open source. Improve this page.