openextract 0.3.1

Extract structured data from documents, images, audio, and video using LLMs

openextract

Extract structured data from documents, images, audio, and video using LLMs.

Installation

uv add openextract

Or

pip install openextract

Usage

from pydantic import BaseModel
from openextract import extract

class PdfInfo(BaseModel):
    summary: str
    language: str

result = extract(
    schema=PdfInfo,
    model="openai:gpt-5.4",
    input_file="https://example.com/document.pdf",
    instructions="return a 2 sentence summary and the primary language of the document",
)
print(result)

Changelog

See CHANGELOG.md for release history.

Contributing

See CONTRIBUTING.md for development setup and guidelines.