cff-version: 1.2.0
message: >-
  If you use py-idp in academic work, please cite it using the metadata
  below. The pipeline architecture (parse → classify → extract → assess
  → validate → HITL) is inspired by the AWS Solutions Library sample
  on accelerated intelligent document processing; PDF parsing wraps
  IBM Docling (arXiv:2408.09869).
type: software
title: "py-idp"
version: 0.6.2
date-released: 2026-09-10
license: AGPL-3.0-or-later
license-url: https://github.com/rollroyces/py-idp/blob/main/LICENSE-AGPL
url: "https://github.com/rollroyces/py-idp"
repository-code: "https://github.com/rollroyces/py-idp"
keywords:
  - intelligent-document-processing
  - ocr
  - document-extraction
  - llm
  - multimodal
  - pydantic
  - rag
  - information-extraction
authors:
  - family-names: "Lam"
    given-names: "Royce"
    alias: "rollroyces"
    email: "roycelam@umich.edu"
preferred-citation:
  type: software
  title: "py-idp: a general-purpose, AI-enabled Intelligent Document Processing framework for Python"
  authors:
    - family-names: "Lam"
      given-names: "Royce"
      alias: "rollroyces"
  year: 2026
  url: "https://github.com/rollroyces/py-idp"
  version: 0.6.2