Anthropic Claude Provider

Use Anthropic's Claude models through SimplerLLM's unified interface with consistent, easy-to-use methods.

Overview

SimplerLLM provides seamless integration with Anthropic's Claude API, allowing you to use their models with just a few lines of code. The unified interface handles authentication, request formatting, and response parsing automatically.

Using Claude Models

Anthropic regularly releases new models and updates. You can use any Claude model by specifying its name in the model_name parameter. SimplerLLM works with all Claude models through the same unified interface.

Model Examples

Here are some commonly used models (as examples):

# Claude 3.5 Sonnet (example)
model_name="claude-3-5-sonnet-20241022"

# Claude 3 Opus (example)
model_name="claude-3-opus-20240229"

# Claude 3 Haiku (example)
model_name="claude-3-haiku-20240307"

# You can use any model name from Anthropic's catalog

Setup and Authentication

To use Anthropic Claude with SimplerLLM, you need an API key from Anthropic:

1. Get Your API Key

2. Configure Environment Variables

Add your API key to a .env file in your project root:

# .env file
ANTHROPIC_API_KEY=sk-ant-your-api-key-here

Basic Usage

Here's how to use Claude models with SimplerLLM:

Simple Text Generation

from SimplerLLM.language.llm import LLM, LLMProvider

# Create Claude LLM instance
llm = LLM.create(
    provider=LLMProvider.ANTHROPIC,
    model_name="claude-3-5-sonnet-20241022"
)

# Generate a response
response = llm.generate_response(
    prompt="Explain the concept of recursion in programming with a simple example"
)

print(response)

With Custom Parameters

llm = LLM.create(
    provider=LLMProvider.ANTHROPIC,
    model_name="claude-3-5-sonnet-20241022",  # Use any Claude model name
    temperature=0.7,      # Controls creativity (0.0 to 1.0)
    max_tokens=1000,      # Maximum response length
    top_p=0.9            # Nucleus sampling parameter
)

response = llm.generate_response(
    prompt="Write a creative short story about AI"
)

print(response)

With System Message

response = llm.generate_response(
    prompt="How do I optimize database queries?",
    system_message="You are an expert database administrator. Provide detailed, technical answers with examples."
)

print(response)

Advanced Features

Structured JSON Output

from pydantic import BaseModel, Field
from SimplerLLM.language.llm_addons import generate_pydantic_json_model

class ProductInfo(BaseModel):
    name: str = Field(description="Product name")
    category: str = Field(description="Product category")
    price: float = Field(description="Price in USD")
    features: list[str] = Field(description="Key features")

llm = LLM.create(provider=LLMProvider.ANTHROPIC, model_name="claude-3-5-sonnet-20241022")

product = generate_pydantic_json_model(
    llm_instance=llm,
    prompt="Generate information for a premium wireless headphone",
    model_class=ProductInfo
)

print(f"Name: {product.name}")
print(f"Category: {product.category}")
print(f"Price: ${product.price}")
print(f"Features: {', '.join(product.features)}")

Conversation with Context

from SimplerLLM.prompts.messages_template import MessagesTemplate

# Create conversation
messages = MessagesTemplate()
messages.add_system_message("You are a helpful Python programming tutor")
messages.add_user_message("What is a list comprehension?")
messages.add_assistant_message("A list comprehension is a concise way to create lists in Python...")
messages.add_user_message("Can you show me an example?")

# Generate response with conversation context
llm = LLM.create(provider=LLMProvider.ANTHROPIC, model_name="claude-3-5-sonnet-20241022")
response = llm.generate_response(messages=messages.messages)

print(response)

Prompt Caching

Anthropic supports prompt caching to reduce costs for repeated prompts:

llm = LLM.create(
    provider=LLMProvider.ANTHROPIC,
    model_name="claude-3-5-sonnet-20241022"
)

# Enable prompt caching for large, repeated context
response = llm.generate_response(
    prompt="Based on the documentation above, how do I implement feature X?",
    prompt_caching=True,
    cached_input="[Large documentation text that will be reused across multiple requests]"
)

print(response)

Configuration Options

Fine-tune Claude model behavior with these parameters:

Pricing Considerations

Anthropic charges based on token usage (both input and output tokens). Different models have different pricing tiers. For current pricing information, visit Anthropic's pricing page.

Best Practices

Error Handling

from SimplerLLM.language.llm import LLM, LLMProvider

try:
    llm = LLM.create(
        provider=LLMProvider.ANTHROPIC,
        model_name="claude-3-5-sonnet-20241022"
    )

    response = llm.generate_response(
        prompt="Your prompt here",
        max_tokens=500
    )

    print(f"Response: {response}")

except ValueError as e:
    # Configuration errors
    print(f"Configuration error: {e}")

except Exception as e:
    # API errors (rate limits, network issues, etc.)
    print(f"API error: {e}")
    # Implement retry logic or fallback here

What's Next?