Skip to content

getting-started.md

Latest commit

 

History

History
133 lines (88 loc) · 3.32 KB

getting-started.md

File metadata and controls

133 lines (88 loc) · 3.32 KB

Getting Started

Home > Docs > Getting Started

Prerequisites

Docling SDK supports three client modes, each with its own requirements:

Client Runtime Requirement
API Node.js >= 18, Bun, Deno A running Docling Serve instance
CLI Node.js >= 18 Python with Docling installed
Web Browser with WebGPU or WASM Peer dependencies (see below)

Installation

npm (recommended)

npm install docling-sdk

GitHub Package Registry

npm install @btwld/docling-sdk

Both packages are identical. The GitHub Package Registry version is available for enterprise environments.

Peer dependencies

All peer dependencies are optional and only needed for specific features:

# Web OCR (browser-based OCR)
npm install @huggingface/transformers onnxruntime-web

# PDF rendering in Web client
npm install unpdf

# TypeScript (development)
npm install -D typescript

Choosing Your Client

Use case Client Why
Server-side conversion through Docling Serve API Fastest, supports streaming, chunking, async tasks
Local conversion without a server CLI Uses Python Docling directly, batch and watch modes
In-browser OCR without any server Web Runs entirely client-side via WebGPU/WASM

First Conversion

API Client

Start a Docling Serve instance, then:

import { readFile } from "node:fs/promises";
import { Docling } from "docling-sdk";

const client = new Docling({
  api: { baseUrl: "http://localhost:5001" },
});

const buffer = await readFile("./document.pdf");
const result = await client.convert(buffer, "document.pdf", {
  to_formats: ["md"],
});

console.log(result.document.md_content);

See the full API Client guide for conversion options, streaming, async tasks, and more.

CLI Client

Install the Python Docling CLI (pip install docling), then:

import { Docling } from "docling-sdk";

const client = new Docling({
  cli: { outputDir: "./output" },
});

const result = await client.convert(
  "./document.pdf",
  "document.pdf",
  { to_formats: ["md"] }
);

console.log(result.document.md_content);

See the full CLI Client guide for batch processing, directory watching, and error handling.

Web OCR

Install the required peer dependencies first:

npm install @huggingface/transformers onnxruntime-web
import { createWebClient } from "docling-sdk/web";

const client = createWebClient({ device: "webgpu" });

client.on("loading", ({ progress, status }) => {
  console.log(`Loading: ${Math.round(progress * 100)}% - ${status}`);
});

await client.initialize();

const result = await client.processImage(imageFile);
console.log(result.markdown);

client.destroy();

The model (~500 MB) is cached in IndexedDB after the first download.

See the full Web Client guide for events, cache management, and standalone utilities.

Next Steps