Instructor

What it does

Instructor is a developer library for turning language-model responses into typed data. Developers define a schema, such as a Pydantic model, provide messages and a model provider, then use Instructor to request a response in that structure. It can validate the result and reask the model after validation errors. It also supports nested schemas and streaming partial results and lists.

The documented workflow is for developers building LLM-backed applications, rather than end users working in a hosted app. The site lists Python, TypeScript, Go, Ruby, Elixir and Rust offerings; Python installation instructions require Python 3.9 or newer. Instructor provides a common interface for multiple model providers, including local-model options such as Ollama. The selected provider or runtime performs inference, so availability and any usage costs depend on that choice. Schema validation and retries do not establish that extracted information is factually correct.

What it can do

• Extract structured data using developer-defined schemas
• Validate outputs with Pydantic
• Reask or retry after validation failures
• Handle nested data structures
• Stream partial results and lists
• Use a common interface across supported LLM providers

Works with

• Pydantic
• OpenAI
• Anthropic
• Google Gemini
• Ollama
• DeepSeek
• Mistral
• Cohere
• llama-cpp-python
• vLLM

Before you choose

• Schema validation does not establish that model-generated data is factually correct.
• The selected provider or local runtime performs inference; access requirements and any associated costs depend on that choice.
• Reasking can make additional model requests; retry limits are configurable.

Listed under

• Coding
• Data extraction
• Developers
• Developer tools
• LLM
• LLM extraction
• Open source

Get started

Python installation command: `pip install instructor`.

1. Install the Python package with `pip install instructor`; the site also documents `uv add instructor` and `poetry add instructor`.
2. Define the desired output schema, for example with a Pydantic model.
3. Select a supported model provider and submit messages with the schema as the response model.

MIT-licensed library; inference through your chosen provider or local runtime is separate.

Source: github.com/instructor-ai/instructor/blob/main/LICENSE, checked 1 Nov 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 1 Nov 2026. Each source supports part of the overview; availability and prices can change.

• Instructor - Multi-Language Library for Structured LLM Outputs | Python, TypeScript, Go, Ruby - Instructor ↗python.useinstructor.com
• Installing Instructor with Pip - Instructor ↗python.useinstructor.com
• instructor/LICENSE at main · 567-labs/instructor · GitHub ↗github.com
• Enhancing AI Validations with Pydantic's Framework - Instructor ↗python.useinstructor.com
• Instructor - Multi-Language Library for Structured LLM Outputs | Python, TypeScript, Go, Ruby - Instructor ↗python.useinstructor.com

Company website

python.useinstructor.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

manual_review

Last checked 1 Nov 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.