DoclingSharp

DoclingSharp is a .NET 8 client library for interacting with Docling (version >1.0.1). It provides two core features:

• DoclingClient – Extracts content from documents uploaded as IFormFile objects via the Docling API.
• DocumentChunker – for splitting document content into manageable chunks suitable for embeddings or other processing.

---

Installation

Install via NuGet:

dotnet add package DoclingSharp

Usage

In your Program.cs or Startup.cs:

builder.Services.AddDocling(c =>
{
    c.DoclingAddress = new Uri("http://localhost:5001");
    c.ChunkMaxCharacters = 2048;
    c.ChunkCharacterOverlap = 128;
});

This sets up both DoclingClient and DocumentChunker with your preferred configuration.

Extract and chunk documents

In a controller, service, or background task:

// Extract content from an uploaded IFormFile
var result = await _doclingClient.ExtractDocumentContentAsync(file);

// Split the content into chunks for embedding or further processing
var documentChunks = _documentChunker.ChunkDocument(result.Markdown);

// Pass the chunks to your embedding engine

Features

• ExtractDocumentContentAsync: Works with IFormFile inputs for PDFs, DOCX, and other supported formats. Returns structured DoclingResult including Markdown, plain text, JSON, and HTML.
• DocumentChunker: Splits documents into overlapping chunks to ensure no content is lost between segments, making it ideal for embedding workflows.
• Fully integrated with IHttpClientFactory and IOptions for clean dependency injection.
• Configurable chunk sizes and overlap.

Example

var options = new DoclingOptions
{
    DoclingAddress = new Uri("http://localhost:5001"),
    ChunkMaxCharacters = 2048,
    ChunkCharacterOverlap = 128
};

var httpClientFactory = ...; // get via DI or manually
var doclingClient = new DoclingClient(httpClientFactory, Options.Create(options));
var documentChunker = new DocumentChunker(options);

var file = File.OpenRead("example.pdf");
var result = await doclingClient.ExtractDocumentContentAsync(file);
var chunks = documentChunker.ChunkDocument(result.Markdown);

Requirements

• .NET 8+
• Docling API >1.0.1
• Microsoft.Extensions.Http and Microsoft.Extensions.Options