Anthropic Cookbook — Official Claude Recipes

What it is

The Anthropic Cookbook is the official repository of notebooks and code recipes for building applications with Claude. Each recipe is a self-contained Jupyter notebook that demonstrates a specific pattern: prompt engineering, tool use, retrieval-augmented generation, multi-step agents, multimodal inputs, and enterprise integration patterns.

The cookbook targets developers who are building production applications with the Claude API. It is maintained by Anthropic and reflects current best practices for the Claude model family.

How it saves time or tokens

Instead of piecing together techniques from blog posts and forum threads, the cookbook gives you tested, copy-paste-ready patterns. Each notebook includes the full code, expected outputs, and explanations of why each approach works. The estimated token cost for a typical recipe is around 500 tokens.

The recipes also demonstrate token-efficient patterns like prompt caching, structured output formatting, and context window management -- directly applicable to reducing API costs.

How to use

1. Clone the repository:

git clone https://github.com/anthropics/anthropic-cookbook.git
cd anthropic-cookbook
pip install anthropic

1. Browse the notebooks by topic. Each is self-contained with its own dependencies.

1. Open any notebook in Jupyter and run the cells:

jupyter notebook misc/prompt_caching.ipynb

Example

# From the tool use cookbook recipe
import anthropic

client = anthropic.Anthropic()

response = client.messages.create(
    model='claude-sonnet-4-20250514',
    max_tokens=1024,
    tools=[{
        'name': 'get_weather',
        'description': 'Get current weather for a location',
        'input_schema': {
            'type': 'object',
            'properties': {
                'location': {'type': 'string'}
            },
            'required': ['location']
        }
    }],
    messages=[{
        'role': 'user',
        'content': 'What is the weather in San Francisco?'
    }]
)

Related on TokRepo

• AI Tools for Coding -- More AI development tools and frameworks
• Prompt Library -- Curated prompt templates for various AI tasks

Common pitfalls

• Some notebooks require specific API features (like tool use or vision) that need the correct Claude model version. Check the model parameter in each notebook.
• The cookbook evolves with Claude's capabilities. Older notebooks may reference deprecated parameters or model names. Always check the last-modified date.
• Running all notebooks sequentially can incur significant API costs. Run them individually and monitor your usage dashboard.