What is Docling?

Docling is an open source document intelligence platform developed by the Linux Foundation that helps teams turn documents into structured, AI-ready data formats. The toolkit simplifies how documents are prepared for search, RAG, and agentic workflows, while preserving structure and context essential for high-quality information retrieval and interpretation by AI.

Why Docling for IBM watsonx?

Enterprise knowledge is trapped in PDFs, images, slide decks, and other data formats that AI cannot reliably use. Utilizing frontier models to interpret this content directly is expensive at scale, while manual preparation is too time-consuming to sustain.

Docling is able to transform these complex data formats into clean, structured data optimized for AI interpretation, reducing the cost and time required to prepare data for AI-powered workflows on IBM watsonx. Docling for IBM watsonx gives users all the capabilities of the open source Docling tool without users needing to host the infrastructure on their own machines.

How is this different from the open-source Docling?

Docling for IBM watsonx is a managed service that eliminates the need to deploy and maintain infrastructure. You can either use the RESTful API or the Python SDK for easy integration with Docling for IBM watsonx. Updates are performed automatically, so you always have the latest models and improvements.

The open-source Docling is ideal for local development and experimentation, while Docling for IBM watsonx is designed for production deployments.

One tool for complex content

You can process PDFs, images, office files, and more through a single document processing approach, reducing tool sprawl and simplifying how content moves into AI workflows.

Preservation of document structure and context

Docling retains document layout, tables, formulas, reading order, and relationships that many other extraction approaches lose, allowing for improvements in retrieval quality and enabling downstream AI to produce more trustworthy results.

More than basic OCR and generic extraction

Using Docling for IBM watsonx can give you access to specialized models for layout analysis, table recognition, and enterprise enrichments, producing higher-quality structured outputs for AI.

Key Features

• Single Convert Endpoint - Low-latency option for agentic AI flows
• Multi-Document Conversion - Convert more than one document in a single request
• Batch Processing - Convert any number of documents from web URLs or cloud storage for large-scale ingestion
• Multi-format Support - PDFs, images, Office files, and more
• Multiple Output Formats - Markdown, Text, JSON, HTML
• Structure Preservation - Maintains layout, tables, formulas, and reading order
• Specialized Models - Layout analysis, table recognition, and enterprise enrichments

Getting Started

To get started with Docling for IBM watsonx:

1. Sign up and onboard at ibm.com/products/docling.
2. Obtain your Service URL and API Key from the Docling for IBM watsonx client.
3. Follow our Quick Start guide to set up your first API call!