D

AI tool

Docling

Parse PDFs and office files into structure for AI workflows

Docling is an open-source document parser for developers building RAG knowledge bases, batch preprocessing and structured extraction. It converts PDFs and office files into a unified document representation and exports formats including Markdown and JSON; it is not a hosted paper-question-answering chatbot. The code uses the MIT license. Local software has no vendor-imposed free page allowance, but you supply compute, storage and model downloads, and external model services may charge separately. Check each model's license independently. Official platforms include Windows, macOS and Linux, with Python 3.10 or later. Documentation and CLI are primarily English; Chinese OCR needs an appropriate engine and language model, with no guarantee of flawless complex scans. Compared with NotebookLM, its directory value is control over parsing and downstream integration, not a ready-made Q&A interface. Initial execution may download models; fully offline operation requires preparing dependencies and weights. Remote services or third-party plugins change data boundaries. Checked v2.145.0, released October 29, 2026. This site reviewed official README, LICENSE, release, OCR and offline configuration documentation without installing or converting a sample.

Last verified: 2026-11-01