A collection of Jupyter Notebook recipes and tutorials by Anthropic demonstrating how to build applications with the Claude API. It covers classification, retrieval, summarization, tool use, and vision tasks through copy-able code snippets.
Project overview
It provides official, Anthropic-backed code examples that developers can copy directly into their projects to integrate Claude effectively.
Project type
Infrastructure
Deployment
Refer to project documentation
License
MIT
Best for
Developers who want to learn how to integrate Claude with external tools and data sources using Python and Jupyter Notebooks.
Key capabilities
Code examples exploring techniques for text and data classification using Claude.
Guides teaching how to enhance Claude responses with external knowledge.
Guides and code recipes demonstrating techniques for effective text summarization with Claude.
Code recipes for integrating Claude with external tools and functions, including customer service agents, calculators, and SQL queries.
Notebooks covering vision tasks with Claude, such as getting started with images, interpreting charts, and extracting content from forms.
Examples demonstrating how to supplement Claude knowledge with external data sources like Pinecone, Wikipedia, web pages, and Voyage AI.
Limitations and risks
Making the most of the examples requires a Claude API key, resulting in a paid service dependency.
Executing the examples requires an understanding of Python and Jupyter Notebooks.
The system requires external services for inference, relying on Claude Models hosted via API rather than local execution.
Getting started
Sign up and obtain a Claude API key, clone the repository, and open and execute the desired Jupyter Notebook locally. Requires understanding of Python.