# Northwind Accounting Services GmbH
## Invoice INV-2024-0042
Pappelallee 18
10437 Berlin
accounts@northwind.example
DE813529441
**Invoice Date:** 2024-03-15
**Due Date:** 2024-04-14
**Payment Terms:** Net 30
## Bill To
Acme Retail Europe
Finance Team
Nieuwezijds Voorburgwal 21
1012 RC Amsterdam
| Description | Hours | Rate | Amount |
| --- | ---: | ---: | ---: |
| Month-end close automation workshop | 6 | USD 120.00 | USD 720.00 |
| Invoice schema rollout and testing | 4 | USD 120.00 | USD 480.00 |
| Vendor onboarding playbook update | 2 | USD 95.00 | USD 190.00 |
**Subtotal:** USD 1,390.00
**Tax (0%):** USD 0.00
**Total Due:** USD 1,390.00
Please remit payment within 30 days via bank transfer using reference **INV-2024-0042**. IBAN: DE42 1001 0010 0987 6543 21.
Input Preview
Output Preview
# Elena Vasquez
## Senior Software Engineer
Berlin, Germany
elena.vasquez@email.com
+49 170 1234567
github.com/elenavasquez
## Professional Summary
Distributed systems engineer with 8 years of experience across Elixir, Kubernetes, and event-driven platforms. Built internal developer tooling, event-driven data pipelines, and high-throughput APIs for B2B SaaS teams.
## Professional Experience
- TechFlow GmbH — Senior Software Engineer (2022-Present) — led event ingestion platform handling 120M jobs/month
- DataBridge AG — Software Engineer (2018-2022) — migrated reporting from batch scripts to a self-serve data platform
- NordStack Labs — Platform Engineer (2016-2018) — built deployment tooling for regulated cloud workloads
## Selected Achievements
- Cut incident recovery time by 42% through automated failover runbooks
- Reduced p95 request latency from 480ms to 170ms on a multi-tenant API
- Mentored 6 engineers into senior and staff-level platform roles
## Education
M.Sc. Computer Science, Technical University of Munich
## Skills
Elixir, Python, Go, Kubernetes, PostgreSQL, Kafka, Terraform, AWS
When an automation moves from prototype to production, you should not have to rebuild it for every environment. Iteration Layer lets scripts, agents, and n8n workflows call the same European AI workflow runtime.
Input
40+ file formats
Extraction
Document, website and
markdown extraction
Generation
Document, image and
sheet generation
Output
Standardized format
Fits into your existing stack
Native SDKs for TypeScript, Python, and Go. OpenAPI spec for everything else. MCP server for AI agents and Claude Code skills. n8n integration for visual workflows.
EU AI workflow runtime
Run document, image, and file steps through one EU-hosted workflow layer with shared API conventions and billing.
Agent-ready by design
Expose the same document and image actions to MCP tools and Claude Code skills, then reuse the API contract when workflows graduate into scripts or automations.
Verified n8n node
Install the verified Iteration Layer node in n8n, then route documents and generated files through the same provider from visual workflows.
Upload any document via URL or base64 — PDF, Office, EPUB, LaTeX, email, images, public website URLs, and more. Any supported format works in the same endpoint.
02
We parse, OCR, and describe
The document is parsed, scanned pages are run through OCR, and tables are extracted. Image files also receive a natural language description of their visual content.
03
Get clean markdown
Receive a JSON result with the file name, MIME type, and extracted markdown. HTML links are rendered as numbered references, and image files also include a plain-language description field.
Intelligent Parsing
The API automatically selects the best parsing approach for your document. Dense tables, multi-column layouts, and mixed content are handled without any configuration.
Headings, paragraphs, tables, lists, and link references are preserved as clean markdown syntax. Website pages drop common navigation and footer boilerplate before output.
Deep Content Understanding
Images and scanned documents aren't treated as pixel grids to OCR. The API understands what they depict — product photos, charts, diagrams — and returns a natural language description alongside the extracted text.
Built-In OCR
Scanned PDFs and image files are automatically run through OCR. You get readable markdown regardless of whether the source is text or pixels.
All Document Formats
40+ file formats plus public website URLs — PDF, DOCX, PPTX, ODT, ODS, XLSX, EPUB, LaTeX, EML, Jupyter, images, and more — all handled by the same endpoint. No format-specific setup or pre-processing required.
No Model Training
Your documents are never used to train or improve AI models. This is guaranteed for all plans — not gated behind an enterprise contract.
Real-world pipelines, ready to ship
Each recipe chains multiple APIs into a complete workflow. Pick one, tweak it, and deploy — or use it as a starting point for your own pipeline.
Your data is processed on EU-hosted infrastructure and never stored beyond temporary logs. Zero data retention, GDPR-compliant workflows, and a Data Processing Agreement are available for every customer.
Learn more about our security practices
.
EU-hosted core processing
Application and processing infrastructure runs in Europe, with provider-scope ISO 27001 and BSI C5 evidence documented for procurement reviews.
Zero data retention
Customer files and processing results are not stored after the request. Usage logs are retained for 90 days and automatically deleted.
Clear answers for security teams
Give reviewers the answers they need up front: where files are processed, what is retained, which subprocessors are involved, and how AI inputs, outputs, review gates, and audit records move through each workflow.
Our OCR benchmark shows strong extraction accuracy, reliability, and performance across 41 real workflow files, including forms, invoices, scans, tables, charts, and photos.
What file formats are supported?
The API accepts 40+ file formats including PDF, DOCX, PPTX, ODT, ODS, XLSX, EPUB, CSV, TSV, HTML, LaTeX, EML, Jupyter notebooks, and all common image formats. Scanned documents are processed with built-in OCR.
What is the difference between this and Document Extraction?
Document to Markdown runs only the ingestion step — it converts files to clean markdown. Document Extraction builds on this by also applying a schema to extract specific fields as structured JSON. Use Document to Markdown when you want the content itself; use Document Extraction when you want specific named values.
Why does the markdown include an image description?
For image files, the API runs both OCR (to extract any text) and a vision model (to describe the visual content). The description is returned as a separate field so you can use it in your own downstream processing.
How many files can I send per request?
Up to 20 files per request. Each file gets its own result in the response array. The order of results matches the order of the input files.
Is the output suitable for LLMs?
Yes. The markdown format is the same used internally by the Document Extraction API as input to LLM extraction. Tables, structure, content, and numbered link references are preserved in a way that models read reliably.
Is Document to Markdown GDPR-compliant?
Yes. Files are processed on EU infrastructure, handled in memory, and not retained after processing. See our security practices and GDPR and AI Act overview for the compliance context.
Build your first workflow in minutes
Chain our APIs into a workflow you can test with your own data. Free trial credits included.