Parsr is a document parsing and extraction toolchain that transforms PDFs, images, and other document formats into structured data.
The tool addresses the problem of converting unstructured documents into machine-readable formats by applying a sequence of processing modules that clean documents, regenerate hierarchy, and detect semantic elements. It accepts images, PDFs, DOCX files, and EML messages as input and outputs results in JSON, Markdown, CSV, or plain text formats. The processing pipeline can identify headings, tables, lists, tables of contents, page numbers, headers, footers, and links while organizing content into words, lines, and paragraphs.
Parsr suits teams needing to automate document analysis, data entry workflows, or archival processes where structured extraction from scanned or digital documents is required. The tool provides both an API and a graphical interface for document submission and result visualization, plus a Python client for programmatic access. It can be deployed via Docker or installed bare-metal depending on infrastructure preferences.
The project is no longer maintained and security patches are not being applied. The README explicitly recommends considering alternatives such as LiteParse for users requiring a local, actively supported parsing solution.
