Docling Serve on Ubuntu 24.04

Overview

Docling Serve is the HTTP API service around Docling, an open source document conversion toolkit. It converts PDF, DOCX, PPTX, HTML and image documents into well structured Markdown or JSON, using machine learning models that understand page layout and reconstruct tables, so headings, paragraphs, lists and tabular data survive the conversion instead of collapsing into a wall of text. Applications POST a document to the API and receive back structured content ready to index, chunk for retrieval augmented generation, feed to a language model or store, and an interactive playground and OpenAPI docs are built in.

It suits teams building search, retrieval augmented generation and document processing pipelines who need a self hosted, private conversion service they run inside their own cloud account, so sensitive documents are parsed on infrastructure they control rather than sent to a third party service.

Common uses

• A private self hosted API for converting documents to structured Markdown and JSON
• Preparing PDFs and office files for retrieval augmented generation and search indexing
• Extracting tables and layout from documents inside your own cloud account

See it running

Real screenshots taken while testing this image against its deployment guide.