AI Providers

Overview

BabyClaw supports multiple AI providers through a unified configuration system. You can configure multiple providers simultaneously and reference models using the provider:model format.

Provider Configuration

All providers are configured in the ai.providers section of your config file:

{
  "ai": {
    "providers": {
      "anthropic": {
        "apiKey": "sk-ant-xxxxxxxxxxxxx"
      },
      "openai": {
        "apiKey": "sk-xxxxxxxxxxxxx"
      }
    },
    "models": {
      "chat": "anthropic:claude-sonnet-4-20250514"
    }
  }
}

Provider Schema

Each provider requires at minimum an API key:

apiKey

string

required

API key for the provider. Must not be empty or "REPLACE_ME".

baseUrl

string

Custom API endpoint URL. Use this for self-hosted instances or API proxies.

Supported Providers

Anthropic (Claude)

Provider ID: anthropicGet your API key: Anthropic ConsoleExample models:

claude-sonnet-4-20250514 - Latest Sonnet model (recommended)
claude-opus-4-6 - Most capable model
claude-haiku-4-5 - Fast and efficient
claude-4.5-sonnet - Previous generation
claude-3.5-sonnet-20241022
claude-3.5-haiku-20241022

Configuration:

{
  "ai": {
    "providers": {
      "anthropic": {
        "apiKey": "sk-ant-xxxxxxxxxxxxx"
      }
    },
    "models": {
      "chat": "anthropic:claude-sonnet-4-20250514"
    }
  }
}

Custom base URL (for proxies):

{
  "anthropic": {
    "apiKey": "sk-ant-xxxxxxxxxxxxx",
    "baseUrl": "https://api.anthropic.example.com"
  }
}

OpenAI

Provider ID: openaiGet your API key: OpenAI PlatformExample models:

gpt-4o - GPT-4 Optimized
gpt-4o-mini - Fast and cost-effective
gpt-4.1 - Latest GPT-4.1
gpt-4.1-mini - GPT-4.1 Mini
gpt-4.1-nano - Smallest GPT-4.1
o3 - Reasoning model
o3-mini - Fast reasoning
o4-mini - Latest mini model

Configuration:

{
  "ai": {
    "providers": {
      "openai": {
        "apiKey": "sk-xxxxxxxxxxxxx"
      }
    },
    "models": {
      "chat": "openai:gpt-4o",
      "vision": "openai:gpt-4o"
    }
  }
}

Azure OpenAI:

{
  "openai": {
    "apiKey": "your-azure-api-key",
    "baseUrl": "https://your-resource.openai.azure.com/openai/deployments"
  }
}

Google Generative AI

Provider ID: googleGet your API key: Google AI StudioExample models:

gemini-2.5-pro - Most capable Gemini model
gemini-2.5-flash - Fast Gemini
gemini-2.0-flash - Previous generation flash
gemini-2.0-flash-lite - Lightweight model

Configuration:

{
  "ai": {
    "providers": {
      "google": {
        "apiKey": "AIzaSyxxxxxxxxxxxxx"
      }
    },
    "models": {
      "chat": "google:gemini-2.5-pro",
      "vision": "google:gemini-2.5-pro"
    }
  }
}

Mistral AI

Provider ID: mistralGet your API key: Mistral ConsoleExample models:

mistral-large-latest - Most capable model
mistral-medium-latest - Balanced performance
mistral-small-latest - Fast and efficient
codestral-latest - Code-specialized model
pixtral-large-latest - Vision model

Configuration:

{
  "ai": {
    "providers": {
      "mistral": {
        "apiKey": "xxxxxxxxxxxxx"
      }
    },
    "models": {
      "chat": "mistral:mistral-large-latest"
    }
  }
}

xAI (Grok)

Provider ID: xaiGet your API key: xAI ConsoleExample models:

grok-3 - Latest Grok model
grok-3-fast - Faster Grok
grok-3-mini - Smaller Grok
grok-3-mini-fast - Fast mini model

Configuration:

{
  "ai": {
    "providers": {
      "xai": {
        "apiKey": "xai-xxxxxxxxxxxxx"
      }
    },
    "models": {
      "chat": "xai:grok-3"
    }
  }
}

OpenRouter

Provider ID: openrouterGet your API key: OpenRouterOpenRouter provides unified access to many AI models from different providers.Example models:

anthropic/claude-sonnet-4
anthropic/claude-opus-4
anthropic/claude-4.5-sonnet
openai/gpt-4o
openai/gpt-4.1
openai/o3
openai/o4-mini
google/gemini-2.5-pro
google/gemini-2.5-flash
deepseek/deepseek-r1
deepseek/deepseek-chat-v3
meta-llama/llama-4-maverick
meta-llama/llama-4-scout

Configuration:

{
  "ai": {
    "providers": {
      "openrouter": {
        "apiKey": "sk-or-xxxxxxxxxxxxx"
      }
    },
    "models": {
      "chat": "openrouter:anthropic/claude-sonnet-4"
    }
  }
}

OpenRouter uses a different model naming format: provider/model-name within the OpenRouter namespace.

Vercel AI Gateway

Provider ID: gatewayGet your API key: Vercel DashboardVercel AI Gateway provides unified access, caching, and analytics.Example models:

anthropic/claude-sonnet-4-20250514
anthropic/claude-opus-4-6
openai/gpt-4o
openai/gpt-4.1
google/gemini-2.5-pro

Configuration:

{
  "ai": {
    "providers": {
      "gateway": {
        "apiKey": "your-gateway-api-key",
        "baseUrl": "https://your-gateway.vercel.app/api"
      }
    },
    "models": {
      "chat": "gateway:anthropic/claude-sonnet-4-20250514"
    }
  }
}

Model Configuration

After configuring providers, specify which models to use:

ai.models.chat

string

required

Primary model for chat interactions. Format: provider:model-id

ai.models.vision

string

Model for vision/image understanding tasks. Format: provider:model-idIf not specified, falls back to the chat model.

Example

{
  "ai": {
    "models": {
      "chat": "anthropic:claude-sonnet-4-20250514",
      "vision": "google:gemini-2.5-pro"
    }
  }
}

Model Aliases

Create shortcuts for commonly used models:

ai.aliases

object

default:"{}"

Dictionary mapping alias names to full model references.Alias names must match the pattern: /^[a-z0-9_-]+$/

Example

{
  "ai": {
    "providers": {
      "anthropic": { "apiKey": "..." },
      "openai": { "apiKey": "..." }
    },
    "models": {
      "chat": "smart"
    },
    "aliases": {
      "smart": "anthropic:claude-opus-4-6",
      "fast": "openai:gpt-4o-mini",
      "vision": "google:gemini-2.5-pro"
    }
  }
}

You can then reference models by alias in your configuration or when switching models.

Multi-Provider Setup

Configure multiple providers to switch between them easily:

{
  "ai": {
    "providers": {
      "anthropic": {
        "apiKey": "sk-ant-xxxxxxxxxxxxx"
      },
      "openai": {
        "apiKey": "sk-xxxxxxxxxxxxx"
      },
      "google": {
        "apiKey": "AIzaSyxxxxxxxxxxxxx"
      },
      "openrouter": {
        "apiKey": "sk-or-xxxxxxxxxxxxx"
      }
    },
    "models": {
      "chat": "anthropic:claude-sonnet-4-20250514",
      "vision": "google:gemini-2.5-pro"
    },
    "aliases": {
      "fast": "openai:gpt-4o-mini",
      "smart": "anthropic:claude-opus-4-6",
      "deepthink": "openai:o3",
      "backup": "openrouter:anthropic/claude-sonnet-4"
    }
  }
}

Environment Variables

While API keys must be configured in babyclaw.json, you can override the config file location:

export BABYCLAW_CONFIG_PATH=/path/to/config.json

Do not commit your configuration file with API keys to version control. Keep it secure and local.

Legacy Configuration

If you’re upgrading from an older version that used ai.gatewayApiKey, you’ll see an error:

Legacy configuration detected.
The 'ai.gatewayApiKey' field has been replaced by 'ai.providers'.
Run 'babyclaw model configure' to migrate.

The new multi-provider system is more flexible and allows you to use multiple AI providers simultaneously.

Validation

Common configuration errors:

Missing API key: Ensure apiKey is present and not "REPLACE_ME"
Invalid provider: Provider ID must be one of: anthropic, openai, google, mistral, xai, openrouter, gateway
Invalid model reference: Must use format provider:model-id
No providers configured: At least one provider must be configured
Invalid base URL: Must be a valid HTTP/HTTPS URL

Best Practices

Start with one provider - Configure Anthropic or OpenAI first, add others as needed
Use aliases - Create memorable shortcuts for frequently used models
Configure vision models - Specify a vision-capable model if you need image understanding
Secure your keys - Never commit API keys to version control
Test configuration - Run babyclaw after configuration changes to validate

Getting API Keys

Provider	Sign Up Link	Free Tier
Anthropic	console.anthropic.com	$5 credit
OpenAI	platform.openai.com	Limited trial
Google	makersuite.google.com	Free quota
Mistral	console.mistral.ai	€5 credit
xAI	console.x.ai	Limited trial
OpenRouter	openrouter.ai	Free credits

Get Started

Core Features

Configuration

Guides

Overview

Provider Configuration

Provider Schema

Supported Providers

Anthropic (Claude)

OpenAI

Google Generative AI

Mistral AI

xAI (Grok)

OpenRouter

Vercel AI Gateway

Model Configuration

Example

Model Aliases

Example

Multi-Provider Setup

Environment Variables

Legacy Configuration

Validation

Best Practices

Getting API Keys

Build docs developers (and LLMs) love

Get Started

Core Features

Configuration

Guides

​Overview

​Provider Configuration

​Provider Schema

​Supported Providers

​Anthropic (Claude)

​OpenAI

​Google Generative AI

​Mistral AI

​xAI (Grok)

​OpenRouter

​Vercel AI Gateway

​Model Configuration

​Example

​Model Aliases

​Example

​Multi-Provider Setup

​Environment Variables

​Legacy Configuration

​Validation

​Best Practices

​Getting API Keys

Build docs developers (and LLMs) love

Overview

Provider Configuration

Provider Schema

Supported Providers

Anthropic (Claude)

OpenAI

Google Generative AI

Mistral AI

xAI (Grok)

OpenRouter

Vercel AI Gateway

Model Configuration

Example

Model Aliases

Example

Multi-Provider Setup

Environment Variables

Legacy Configuration

Validation

Best Practices

Getting API Keys