Autype Lens
#AI-powered document understanding
Autype Lens goes beyond plain OCR. It combines optical character recognition with vision-language models to extract text, layout, styling, and structure from any document.
Document Input
PDF, DOCX, ODT
Structured Output
md / mdd / json
Document Input
PDF, DOCX, ODT
Structured Output
md / mdd / json
Two AI technologies, one pipeline
Traditional OCR captures text but loses everything else. Vision-language models understand pages visually but struggle with precision. Autype Lens combines both into a single pipeline (built on fine-tuned open-source models) to deliver results neither approach could achieve alone.
Optical Character Recognition
Precision text extraction at character level. Our fine-tuned OCR models handle complex layouts, multi-column pages, and embedded tables with high accuracy, including scanned documents.
Vision-Language Models
Page-level visual comprehension. Fine-tuned VLMs analyze the full page as an image to understand heading hierarchies, font styles, color schemes, margins, and spatial relationships between elements.
Unified extraction pipeline
The OCR pass delivers precise text. The VLM pass delivers document semantics. Our pipeline merges both, validates output against JSON schemas, and retries automatically if needed.
Four ways to understand documents
Each operation targets a different document understanding task. All are accessible via a single REST API endpoint.
Revenue grew by 23% compared to last quarter, driven by enterprise adoption and new product launches across all regions.
- ●Onboarded 18 new clients
- ●Retention rate at 94%
- ●3 new markets planned for Q1
International expansion is on track with 3 new markets planned for Q1 2025. Operating margins improved to 18%.
# Quarterly Report Q4 Revenue grew by **23%** compared to last quarter, driven by enterprise adoption across all regions. Key achievements this quarter: - Onboarded **18 new clients** - Retention rate at 94% - 3 new markets planned for Q1  ## Key Metrics :::table{}headerBg=#f0f0f0 rowAltBg=#fafafa{} | Metric | Value | Change | |----------|--------|--------| | Revenue | €1.2M | +23% | | Clients | 142 | +18 | | Margin | 18% | +3% | :::
Text, headings, tables, and structure preserved as Markdown or JSON.
Smart OCR
Convert documents to Markdown, Autype extended Markdown (with styling & defaults), or complete Autype JSON. All formats are ready to edit or re-render. Supports page selection.
Service Agreement
This Service Agreement (“Agreement”) is entered into as of March 15, 2025, by and between Sterling & Associates (“Provider”) and Globex Industries (“Client”).
1. Scope of Services
The Provider shall deliver consulting services as described in Exhibit A, including strategic planning, market analysis, and quarterly reviews.
2. Compensation
The Client agrees to pay a monthly retainer of €8,500 for the duration of this Agreement, due within 30 days of invoice.
3. Term & Termination
This Agreement shall commence on the date above and continue for a period of 12 months, unless terminated by either party with 30 days written notice.
{ "fileId": "550e8400-...", "labels": [ "contract", "invoice", "report", "letter" ] }
{ "category": "contract", "confidence": 0.95, "labels": [ { "name": "contract", "score": 0.95 }, { "name": "letter", "score": 0.03 }, { "name": "report", "score": 0.02 } ] }
Returns the matched category with a confidence score.
Document classification
Automatically categorize documents into your custom categories. Upload a document, provide your labels, and get back the best match with a confidence score.
Acme Corp
123 Business St, Munich
INVOICE
INV-2025-0042
Bill to:
Globex Industries
Berlin, Germany
Date: 2025-03-15
Due: 2025-04-15
All amounts in EUR
Payment terms: 30 days net · IBAN: DE89 3704 0044 0532 0130 00
{ "fileId": "550e8400-...", "schema": { "invoiceNumber": "string", "date": "string", "vendor": "string", "total": "number", "currency": "string", "lineItems": "array" } }
{ "invoiceNumber": "INV-2025-0042", "date": "2025-03-15", "vendor": "Acme Corp", "total": 4250.00, "currency": "EUR", "lineItems": [ { "description": "Consulting", "amount": 3500 }, { "description": "Expenses", "amount": 750 } ] }
Structured JSON with every field from your schema filled in.
Structured data extraction
Define a field schema and let Lens pull structured data from any document. Works for invoice numbers, dates, names, amounts, and anything else your workflow needs.
Acme Corp
123 Business St, Munich
INVOICE
INV-2025-0042
Bill to:
Globex Industries
Berlin, Germany
Date: 2025-03-15
Due: 2025-04-15
Payment terms: 30 days net · IBAN: DE89 3704 0044 0532 0130 00
{ "fileId": "550e8400-...", "filenameSchema": "invoice-{number}_{date}" }
[ { "fileId": "550e8400-...", "filename": "invoice-INV-2025-0042_2025-03-15" }, { "fileId": "7a3b1c90-...", "filename": "invoice-INV-2025-0039_2025-03-10" }, { "fileId": "b2f4e8d1-...", "filename": "invoice-INV-2025-0038_2025-03-08" }, { "fileId": "c9d5a6f2-...", "filename": "invoice-INV-2025-0035_2025-03-01" }, { "fileId": "e1a7b3c4-...", "filename": "invoice-INV-2025-0033_2025-02-28" }, { "fileId": "f8c2d9e5-...", "filename": "invoice-INV-2025-0031_2025-02-25" }, { "fileId": "a4b6c8d0-...", "filename": "invoice-INV-2025-0029_2025-02-20" }, { "fileId": "d3e5f7a9-...", "filename": "invoice-INV-2025-0027_2025-02-15" }, { "fileId": "1b2c3d4e-...", "filename": "invoice-INV-2025-0024_2025-02-10" }, { "fileId": "5f6a7b8c-...", "filename": "invoice-INV-2025-0021_2025-02-05" }, ... // 25 more results ]
Pattern placeholders replaced with values read from the document.
AI filename generation
Provide a naming pattern like invoice-{number}-{date} and Lens reads the document to fill in the placeholders. Automate your file organization.
Images included in extraction
When using mdd or JSON output, Lens detects and extracts all embedded images from the document. Every image gets a download URL so you can use them in your pipeline or re-render them in a new document. No extra step required.
Three output levels
Choose the depth of extraction you need. From raw text to a fully styled, renderable document.
Standard Markdown
Raw text extraction as clean Markdown. Fast, lightweight, and ideal for search indexing or content migration.
# Quarterly Report Revenue grew by **23%** compared to last quarter. ## Key Metrics | Metric | Value | |----------|-------| | Revenue | €1.2M | | Growth | 23% |
Autype Extended Markdown
Markdown plus full document settings, styling defaults, headers, and footers. Re-render with the original look.
---document size: A4 marginTop: 2.5 marginBottom: 2 ---defaults fontFamily: Inter fontSize: 11 color: #333333 --- # Quarterly Report Revenue grew by **23%**
Autype Document JSON
Complete structured document with sections, elements, and styling. Ready to import into Autype or process programmatically.
{ "defaults": { "fontFamily": "Inter", "fontSize": 11 }, "sections": [{ "type": "flow", "content": [ { "type": "h1", "text": "Report" } ] }] }
Built for real workflows
Autype Lens fits into document pipelines where raw OCR isn't enough.
Document digitization
Convert scanned PDFs and legacy documents into editable, structured formats. The original layout and design are preserved.
Content migration
Move documents between systems without losing formatting. Extract styled content and re-render it in Autype or any other platform.
Automated filing
Classify incoming documents, extract key fields, and generate filenames automatically. Build hands-free document intake pipelines.
Data extraction
Pull structured data from invoices, contracts, reports, and forms. Define your schema once and extract at scale via the API.
One API call away
Integrate Lens into your workflow with a simple REST API call. Upload a file, choose your output format, and get structured results.
curl -X POST https://api.autype.com/api/v1/dev/tools/lens/ocr \ -H "X-API-Key: your_api_key" \ -H "Content-Type: application/json" \ -d '{ "fileId": "550e8400-e29b-41d4-a716-446655440000", "outputFormat": "mdd" }'
{ "id": "job_abc123", "status": "COMPLETED", "result": { "outputFormat": "mdd", "content": "---document\nsize: A4\n---defaults\nfontFamily: Inter\n---\n# Report Title\n\nContent with **styling** preserved..." } }
Available on all plans. 4 credits per page.
Ready to go beyond OCR?
Start extracting text, layout, and styling from your documents today.