Docling Serve icon

Docling Serve

Convert documents using Docling Serve API

Actions17

Chunk → Chunk From URL (Async)

AI-generated

Summary

Chunk a document located at a given HTTP URL asynchronously using the Docling Serve API by submitting a chunking task, polling until completion, and then retrieving the chunked document result.

Inputs

  • Chunker Type — Choose the chunking strategy: 'Hybrid' for multiple strategies combined, or 'Hierarchical' based on document structure.
  • Source URL (required) — Provide the HTTP URL of the document (e.g., a PDF) to chunk asynchronously.
  • Additional Options — Configure advanced chunking and document conversion parameters such as OCR processing (enable or force OCR), table structure extraction, image handling including export mode and scale, text inclusion options, max tokens per chunk, pipeline selection, tokenizer model (only for Hybrid chunker), and processing timeouts. These options customize the chunking behavior.

Output shape

a single JSON object representing the fully processed chunked document retrieved after asynchronous task completion

This operation initiates an asynchronous chunking task and polls the Docling Serve API until the task completes or fails. On success, it returns the complete chunked document data in JSON format paired with the input item. If the task fails or times out, an error is thrown.

Examples

Example 1: Asynchronously chunk a PDF document accessible at a URL using the Hybrid chunker type, with default OCR and image extraction enabled.

Set 'Chunker Type' to 'Hybrid', specify the document's 'Source URL', optionally adjust 'Additional Options' as needed, and run the operation to receive chunked document data after completion.

Links

Discussion