The Three-Service Pipeline Problem
You need to process invoices in an n8n workflow. So you wire up Mistral OCR to extract the text, GPT-4o to structure it into JSON, and a Google Sheets node to store the results. Three services, three sets of credentials, three billing accounts. The OCR costs per page, the LLM costs per token, and every run burns credits on both.
Then someone asks you to generate a PDF summary from the extracted data. Now you need a fourth service — maybe PDF.co, maybe an HTML-to-PDF node, maybe a Function node that shells out to Puppeteer. Each option has its own quirks, its own auth, and its own failure modes.
This is how most n8n document workflows end up. Not because the approach is wrong, but because no single tool covers the full pipeline. Extraction needs one vendor. Transformation needs another. Generation needs a third. The workflow becomes a patchwork of HTTP Request nodes, Function nodes for format conversion, and base64 encoding/decoding glue.
One Node, Full Pipeline
The Iteration Layer n8n community node covers extraction, transformation, and generation in a single install. Pick a resource from the dropdown — Document Extraction, Image Transformation, Image Generation, Document Generation, Sheet Generation, or Document to Markdown — fill in the parameters, and the node handles auth, binary data, and response parsing.
The practical difference: a pipeline that needed three separate services and five n8n nodes becomes two Iteration Layer nodes wired together. Binary data flows between them natively. No Function nodes to decode base64. No format conversion glue.
One API key, one credit pool, and every resource’s output is designed to feed into the next.
What Makes This Different
Most n8n document processing nodes solve one piece of the puzzle. PDF.co handles PDF conversion. Tesseract handles OCR. DALL-E handles image generation. Each is good at its specific job, but chaining them together means managing multiple integrations, multiple billing systems, and multiple failure modes.
Iteration Layer is built as a pipeline from the start. Extract data from a document, transform an image from the extracted content, generate a branded report, export it as an Excel spreadsheet — all through the same node, the same credentials, and the same credit pool.
The extraction is deterministic and schema-based. You define the fields you want, their types, and their descriptions. The API returns structured JSON with confidence scores, citations, and source references. No per-token LLM billing surprises.
Installation
Open Settings > Community Nodes in the n8n UI. Click Install a community node, enter n8n-nodes-iterationlayer, and install.
For self-hosted n8n:
cd ~/.n8n
npm install n8n-nodes-iterationlayerRestart n8n after installation. Go to Credentials > New Credential, search for Iteration Layer API, and enter your API key. Get one at platform.iterationlayer.com — free trial credits included, no credit card required.
Example: Invoice to PDF Report
This is the kind of pipeline that normally needs three services. With Iteration Layer, it’s two nodes.
Node 1: Iteration Layer (Document Extraction) — Set the resource to Document Extraction, feed it the invoice PDF (from an email trigger, a file upload, or a URL), and define an extraction schema with fields for invoice number, vendor name, line items, and total. The node returns structured JSON with confidence scores on every field.
Node 2: Iteration Layer (Document Generation) — Set the resource to Document Generation, pick PDF as the output format, and build the document definition using the extracted data via n8n expressions. The node returns binary PDF data that plugs directly into a Send Email node, a Google Drive upload, or an S3 write.
Two nodes. One set of credentials. No OCR service, no LLM service, no PDF generation service. The extraction output feeds directly into the generation input.
Example: Product Image Pipeline
Node 1: Iteration Layer (Image Transformation) — Set the resource to Image Transformation, pick up the image from a trigger or file node, and add operations: resize to 1200x1200, smart crop, sharpen, convert to WebP. All 24 operations are configurable through the n8n UI — no JSON editing. The node returns binary image data.
Node 2: Iteration Layer (Image Generation) — Set the resource to Image Generation, define a canvas with the processed image as a layer, add text layers for product name and price, and output as PNG. The node returns the composed image as binary data, ready for upload.
One API call handles the transformation. One handles the composition. The binary output of the first plugs into the second without any conversion step.
Beyond These Two
The same pattern applies to every combination of Iteration Layer resources:
- Document to Markdown → Document Extraction — convert a complex PDF to clean Markdown, then extract structured data from it
- Document Extraction → Sheet Generation — extract data from a batch of invoices and generate a formatted Excel report
- Image Transformation → Document Generation — process product photos and compile them into a PDF catalog
Every resource accepts binary data from a previous node or a URL. Every resource that produces files outputs n8n binary data. The outputs chain into downstream nodes — S3, Google Drive, Slack, email — without intermediate steps.
Get Started
Install the community node from the n8n UI and grab a free API key at platform.iterationlayer.com. The n8n integration docs cover every resource and parameter in detail.
If you’re currently using HTTP Request nodes to call document or image APIs, the community node replaces all of that glue. If you’re chaining multiple services for extraction + generation, the node consolidates them under one integration. And if you’re new to document and image automation in n8n, this is the fastest way to get a pipeline running.