Easy File Upload
Upload documents directly via multipart form. Support for PDF, DOCX, XLSX, and many more formats.
- NV-24187.pdf212 KB
- q3-expenses.xlsx48 KB
- service-agreement.docx96 KB
- receipt-scan.jpg1.4 MB
Everything you need for structured data extraction
A complete API for converting documents and extracting structured data. No ML expertise required. Just define a schema and go.
Document Conversion
Convert PDFs, Word docs, spreadsheets, and more to clean Markdown. Preserves tables, links, and formatting.
NV-24187.md
## Line items| Description | Qty | Amount || --- | ---: | ---: || Carbide end mill 12 mm | 24 | 11 664,00 || ER32 collet set | 3 | 6 435,00 || Coolant concentrate 20 L | 6 | 5 475,00 |Schema-Driven Extraction
Define a JSON schema, get structured data back. Context-aware extraction that works across document types.
schema.json
{
"type": "object",
"properties": {
"invoice_number": { "type": "string" },
"issue_date": { "type": "string" },
"total": { "type": "number" }
}
}output.json
{
"invoice_number": "NV-24187",
"issue_date": "2026-09-14",
"total": 29467.50
}Extraction
Pipeline Processing
Convert and extract in a single API call. Upload a PDF, define a schema, get JSON back. No intermediate steps.
OCR Support
Extract text from scanned documents and images. Perfect for invoices, receipts, and legacy documents.
JSON Schema Validation
Define your output structure with JSON Schema. Get predictable, validated data every time.
Operations
Readable Results & Exports
Review fields and tables in the Playground, then download Excel, CSV or JSON. Read available document text and save it as Markdown. Works for single documents, batches and history.
Budget Alerts
Set monthly spending limits and get notified at configurable thresholds. Never get surprised by bills.
Secure by Design
API key authentication, encrypted data in transit, and automatic data retention policies. Your documents stay safe.
Document extraction in action
Upload a document, get structured JSON back
NV-24187.pdf
Faktura
Industrigatan 14, 417 07 Göteborg
Bill to
Kvarnby Precision AB
Stenbärsgatan 3, 212 36 Malmö
- Invoice no.
- Date
- Due
| Description | Qty | Unit | Amount |
|---|---|---|---|
| 486,00 | |||
| 2 145,00 | |||
| 912,50 |
- Subtotal
- VAT 25 %
- Total
VAT reg. SE556843129101. Bankgiro 5832-1197. Payment within 30 days.
application/json
{
"supplier": ,
"invoice_number": ,
"issue_date": ,
"due_date": ,
"currency": ,
"line_items": [
{
"description": ,
"quantity": ,
"amount":
},
{
"description": ,
"quantity": ,
"amount":
},
{
"description": ,
"quantity": ,
"amount":
}
],
"subtotal": ,
"vat_amount": ,
"total":
}Explore use cases
- Invoice & Receipt ExtractionExtract line items, totals, dates, and vendor details from invoices automatically. Replace manual data entry with structured JSON output.
- Legacy PDF DigitizationConvert scanned documents and legacy PDFs into structured, searchable data. Unlock value from previously inaccessible archives.
- Contract & Agreement ParsingPull key clauses, dates, parties, and obligations from legal documents. Consistent results across hundreds of contracts.
- Internal Document AutomationAutomate processing of HR forms, reports, and internal documents at scale. Eliminate manual data entry across your organization.
Less setup. Your schema. One document price.
Compare Smole with Amazon Textract and Google Document AI. Go from mixed document formats to application-ready JSON in one workflow.
Swipe the table sideways to compare all three
| Feature | Smole | Amazon Textract | Google Document AI |
|---|---|---|---|
| Getting started | Create an API key, upload a file and choose a schema | Configure an AWS account and access permissionsSource | Configure a cloud project, billing, permissions and a processorSource |
| Output for your application | JSON validated against your JSON Schema | Blocks, key-value pairs and query answers to map into your appSource | A Document object with extracted entities to map into your appSource |
| Word, Excel and HTML files | Convert and extract through the same pipeline | Convert first: accepts PDF, JPEG, PNG and TIFFSource | Office and HTML support is limited to Layout ParserSource |
| Custom fields | Define a reusable JSON Schema | Ask Queries or train a Custom Queries adapterSource | Define fields in a Custom Extractor processorSource |
| How usage is priced | Per document: one uploaded file, regardless of pages | Per page, with rates by analysis featureSource | Per page, with rates by processorSource |
Based on official documentation reviewed September 11, 2026. Compares the named services; capabilities vary by API and processor. All three support OCR and structured extraction. Smole pricing follows your plan; technical upload limits apply.
See Smole plans and document pricingReliable document processing at scale
- Handles scanned + digital PDFs
- OCR-powered conversion for scanned documents, native text extraction for digital files.
- Scales to large batches
- Process hundreds or thousands of documents through the same API. Built for batch workloads.
- Schema-driven output
- You define the JSON structure. Every extraction follows your schema for consistent, predictable results.
- Deterministic output structure
- Same schema, same document type, same output structure. No surprises in production.
Ready to automate document processing?
Start extracting structured data from your documents in minutes.