Instructor Review — Joute

The essentials in 20 seconds

Python library that wraps LLM calls to guarantee structured outputs validated by Pydantic
Automatic retry if the LLM output doesn't match the expected schema
Compatible with OpenAI, Anthropic, Google, and any OpenAI-compatible LLM
Price: free, open source MIT

Verdict: Instructor is the simplest and most reliable way to extract structured data from an LLM. If you use Pydantic (and you should), the integration is natural. Essential.

What is Instructor?

Instructor is a Python library created by Jason Liu that radically simplifies structured data extraction from LLMs. You define a Pydantic model, call the LLM, and get back a validated Python object.

If the LLM returns something that doesn't match the schema, Instructor automatically retries the call with the error message in context.

from instructor import patch
from openai import OpenAI
from pydantic import BaseModel

client = patch(OpenAI())

class Person(BaseModel):
    name: str
    age: int

person = client.chat.completions.create(
    model="gpt-4",
    response_model=Person,
    messages=[{"role": "user", "content": "Jason Liu, 30 ans"}]
)
# person.name == "Jason Liu", person.age == 30

Strengths

Native Pydantic validation

Your existing Pydantic models work directly. No need to learn a new schema format.

Smart retry

When the LLM produces invalid output, Instructor reinjects the validation error into the prompt so the LLM can correct itself. The success rate climbs significantly.

Multi-LLM support

OpenAI, Anthropic (Claude), Google (Gemini), Mistral, and any OpenAI-compatible endpoint.

Limits

Python only

No native JavaScript/TypeScript version (BAML covers this need on the TypeScript side).

Strong Pydantic dependency

If your project doesn't use Pydantic, you need to adopt it. In practice this isn't a limitation since Pydantic has become a standard.

Pricing

Free, open source MIT

Alternatives

BAML for TypeScript-first structured extraction
DSPy for a more complete approach to LLM pipeline programming
Outlines for structured generation with local models

Verdict

Instructor is one of the most useful libraries in the Python LLM ecosystem. If you're building pipelines that extract data from texts or documents, install Instructor immediately.