Skip to content
Xberg

n8n

The @xberg-io/n8n-nodes-xberg package is a community node that runs Xberg document extraction inside your n8n workflows. It uses the native @xberg-io/xberg binding, so extraction happens in-process — no API key, no external call, and no data leaves your n8n instance.

npm License

Because it ships a native N-API addon, the node runs on self-hosted n8n only — n8n Cloud does not load native addons, so this package is not among the Cloud-verified nodes.

How it works

Section titled “How it works”
  1. Feed — A workflow passes an uploaded file (binary) or a URL/path to the node.
  2. Extract — Xberg parses the document in-process, running OCR where needed.
  3. Emit — The node returns extracted content, metadata, tables, and optional chunks as item JSON, ready for the rest of your workflow.

Installation

Section titled “Installation”

In your self-hosted n8n instance, open Settings → Community Nodes → Install, enter the package name, and confirm you accept the risks of running community code:

@xberg-io/n8n-nodes-xberg

Or install it manually:

Terminal window
npm install @xberg-io/n8n-nodes-xberg

Node.js 20.15+ is required. Prebuilt binaries ship for Linux (x64/arm64, glibc or musl), macOS (arm64), and Windows (x64/arm64).

Operations

Section titled “Operations”

The node has one resource, Document, with three operations.

Operation Description
Extract Extract text, tables, and metadata from one document per item.
Extract Batch Extract every incoming item in a single batch call — substantially faster than looping.
Map URL List the URLs reachable from a web page or sitemap without extracting them.

Extract and Extract Batch accept either uploaded binary data or a URL/local path via the Input Source parameter, and expose the key ExtractionConfig options: output format (Markdown, plain, HTML, Djot, JSON tree, structured JSON), OCR (enable, force, languages), chunking (size and overlap), image extraction, and quality processing. Per-item failures respect n8n’s Continue On Fail setting.

Output

Section titled “Output”

Each Extract output item’s JSON holds the extracted content (default field text), plus mimeType, extractionMethod, detectedLanguages, and counts (pages, tables, images). metadata, tables, and chunks are attached when their toggles are set, and formattedContent, qualityScore, entities, and summary appear when populated. Enabling Return As Binary also attaches the content as a binary property.

For the full parameter reference and workflow examples, see the package readme. For the list of supported file types, see format support.