ContextQMD
Back to Bentopdf

PDF Workflow Builder

docs/tools/pdf-workflow.md

2.7.05.6 KB
Original Source

PDF Workflow Builder

The Workflow Builder lets you chain multiple PDF operations into a single automated pipeline. Instead of running tools one at a time, you wire up a visual graph of nodes -- input, processing, output -- and execute everything with one click. Your files never leave the browser.

How It Works

The interface is split into three panels: a node toolbox on the left, the canvas in the center, and a settings panel on the right that appears when you select a node.

  1. Add nodes by clicking them in the left sidebar, or drag them directly onto the canvas. Use the search bar at the top of the toolbox to filter by name.
  2. Connect nodes by dragging from an output socket (bottom of a node) to an input socket (top of another node). Connections flow top-to-bottom.
  3. Configure settings by clicking a node to open its settings panel on the right. Each node type has its own options (page ranges, quality sliders, passwords, etc.).
  4. Run the workflow by clicking the Run button in the toolbar. The engine executes nodes in topological order, processing each one in sequence. A status bar at the bottom tracks progress.
  5. Download results -- the terminal Download node saves the final PDF (or ZIP if multiple files are produced).

Node Categories

Nodes are organized into six categories, each color-coded on the canvas:

  • Input (blue) -- PDF Input, Image Input, Word/Excel/PowerPoint to PDF, and 15+ other format converters that feed files into the pipeline.
  • Edit & Annotate (indigo) -- Crop, Greyscale, Invert Colors, Watermark, Page Numbers, Header & Footer, Scanner Effect, Adjust Colors, Background Color, Remove Blank Pages, Remove Annotations.
  • Organize & Manage (violet) -- Merge, Split, Extract Pages, Rotate, Delete Pages, Reverse, Add Blank Page, Divide Pages, N-Up, Fix Page Size, Combine to Single Page, Booklet, Posterize, Edit Metadata, Table of Contents, OCR.
  • Optimize & Repair (amber) -- Compress, Rasterize, Linearize, Deskew, PDF to PDF/A, Font to Outline, Repair.
  • Secure PDF (rose) -- Encrypt, Decrypt, Sanitize, Flatten, Digital Sign, Redact.
  • Output (teal) -- Download (auto-detects single PDF vs. ZIP), PDF to Images, PDF to Text, PDF to DOCX, PDF to XLSX, PDF to CSV, PDF to SVG, PDF to Markdown, Extract Images.

Saving and Loading Workflows

The toolbar provides four persistence options:

  • Save -- stores the current workflow as a named template in your browser's local storage. You can save multiple templates.
  • Load -- opens a modal listing all saved templates. Click Load to restore one, or delete templates you no longer need.
  • Export -- downloads the workflow as a .json file so you can share it or back it up.
  • Import -- loads a .json workflow file from disk, restoring all nodes, positions, and connections.

Templates persist across sessions in local storage. Exported JSON files are portable between browsers and machines.

Execution Model

The engine builds a dependency graph from your connections and executes nodes in topological order. If any node fails, execution halts and the failing node is marked red. A few rules to keep in mind:

  • Every workflow needs at least one Input node and one Output node.
  • The engine detects circular dependencies and blocks execution with an error.
  • The Encrypt node must be the last processing step before the output node. Placing it earlier will trigger a validation warning because downstream nodes cannot operate on encrypted data.
  • Password-protected PDFs prompt for a password during execution.
  • Multi-PDF inputs are supported -- the Merge node can combine them, or batch-processing nodes will process each PDF individually.

Socket Types

Connections use typed sockets to prevent invalid wiring:

  • PDF (indigo) -- a single PDF document.
  • MultiPDF (amber) -- multiple PDF documents bundled together.
  • Image (green) -- an image blob.

Options

Every node has its own configuration accessible through the right settings panel. Common patterns include:

  • Page ranges on Split, Extract Pages, Delete Pages nodes
  • Quality sliders on image conversion output nodes
  • Password fields on Encrypt and Decrypt nodes
  • Text inputs for Watermark, Header & Footer, Edit Metadata
  • Rotation angles on the Rotate node
  • Compression level on the Compress node
  • OCR language selection on the OCR node

Use Cases

  • Invoice processing pipeline: PDF Input -> OCR -> Extract Tables -> Download as CSV.
  • Report standardization: Word to PDF -> Fix Page Size -> Add Page Numbers -> Watermark -> Compress -> Download.
  • Archival preparation: PDF Input -> Repair -> Flatten -> Sanitize -> PDF to PDF/A -> Download.
  • Batch image extraction: PDF Input (multiple files) -> Extract Images -> Download ZIP.
  • Print-ready booklets: PDF Input -> Booklet -> Crop -> Compress -> Download.
  • Secure distribution: PDF Input -> Redact -> Remove Annotations -> Encrypt -> Download.

Tips

  • Use Delete or Backspace to remove the selected node from the canvas.
  • Click the X button on a node's status bar to delete it directly.
  • Pan the canvas by clicking and dragging on empty space. The cursor changes to a grab icon while panning.
  • On mobile, use the + button in the toolbar to toggle the node toolbox.
  • If your workflow produces multiple output files, the Download node automatically bundles them into a ZIP.
  • Start simple -- build and test a two-node pipeline before adding complexity.