Instructor - Multi-Language Library for Structured LLM Outputs

Description

Instructor is a popular library that facilitates the extraction of structured data from large language models (LLMs). It is designed for developers who require reliable and validated outputs, making it an essential tool in the AI landscape. With support for over 15 providers, including OpenAI, Anthropic, Google, and more, Instructor allows users to seamlessly integrate various LLMs into their applications.

Built on Pydantic, Instructor ensures type-safe data extraction with automatic validation and retries. This means that developers can define Pydantic models to specify exactly what data they want from their LLMs, ensuring that the outputs are both structured and validated. The library's automatic retry logic simplifies error handling, allowing users to focus on building their applications rather than managing validation failures.

One of the standout features of Instructor is its support for streaming, which enables real-time processing of partial responses and lists. This capability is particularly beneficial for applications that require immediate feedback or interaction with users. Additionally, Instructor's multi-provider support means that developers can switch between different LLMs without changing their code, making it a versatile choice for various projects.

Instructor is not just limited to Python; it also supports TypeScript, Go, Ruby, Elixir, and Rust, making it accessible to a wide range of developers. The library's simplicity and transparency are key to its user-centric design, allowing for easy integration into existing workflows. For those looking to enhance their projects with quality gates, shareable runs, or built-in observability, PydanticAI, the official agent runtime from the Pydantic team, can be utilized alongside Instructor.

In summary, Instructor is a powerful tool for anyone looking to extract structured data from LLMs efficiently. Its combination of type safety, validation, and support for multiple languages and providers makes it a go-to solution for developers in the AI field.

Instructor's Core Features

• Structured Outputs: Define Pydantic models for data extraction

• Automatic Retries: Built-in retry logic for validation failures

• Data Validation: Leverage Pydantic's validation for response quality

• Streaming Support: Real-time processing of partial responses

• Multi-Provider: Works with 15+ LLM providers

• Type Safety: Full IDE support with type inference

• Open Source Support: Run open source models locally

• Simple API: Complete control over prompts

Getting Started with Instructor

1. 1

   Install Instructor via package manager

2. 2

   Define Pydantic models to specify desired data

3. 3

   Use the from_provider function to extract data from LLMs

4. 4

   Handle validation and retries automatically

5. 5

   Utilize streaming support for real-time processing

6. 6

   Integrate with multiple LLM providers seamlessly

7. 7

   Explore advanced features as needed

8. 8

   Refer to documentation for detailed guidance

Instructor's Use Cases

• Data Extraction
• Real-time Processing
• Multi-Provider Integration
• Error Handling
• Schema Validation

FAQ from Instructor

• What is Instructor?

  Instructor is a Python library for extracting structured, validated data from large language models.

• Which LLM providers does Instructor support?

  Instructor supports over 15 providers including OpenAI, Anthropic, Google, and more.

• Do I need to know Pydantic to use Instructor?

  Basic Pydantic knowledge is helpful, but you can start with simple models.

• How does Instructor handle validation?

  Instructor uses Pydantic's validation to ensure output quality and automatically retries on failures.

• Can I use Instructor with open source models?

  Yes, Instructor works with Ollama, llama-cpp-python, and other local models.

• Does Instructor support async code?

  Yes, Instructor fully supports async/await functionality.

• What programming languages does Instructor support?

  Instructor supports Python, TypeScript, Go, Ruby, Elixir, and Rust.

• How does Instructor compare to other libraries?

  Instructor focuses on structured outputs with automatic validation and retries, excelling in these areas.