1B+ documents processed a year — for 3,000+ teams

Stop parsing documents by hand

One API to extract, generate, convert, and review documents — built so agents can act on the output and people can verify it.

Quote document with extracted data highlighted

Used by Lufthansa, Disney, Autodesk, UBS, Dropbox, IBM
Lufthansa
Disney
Autodesk
UBS
Dropbox
IBM

Try it live

Drop in a document — get structured data back. No signup required.

Switch processing modes and inspect live output from a sample invoice as rendered Markdown, raw Markdown, or spatial JSON.

Start free

5,000 free Data Extraction API credits every month

No credit card required.

Your free plan includes:

    • 5,000 Data Extraction API credits each month
    • Parse and extract from PDFs, images, and Office files
    • Markdown or spatial JSON output
    • Available modes: text, structure, and understand

Benchmarked in the open

Accurate parsing you can reproduce, run after run

See the full benchmark

On opendataloader-bench — 200 real-world PDFs scored with the public DP-Bench metrics — Nutrient's licensed --vision tier tops every accuracy metric, including tables, while running faster than docling. The free default engine leads reading order among the open-source parsers you'd otherwise reach for.

Overall extraction accuracy opendataloader-bench · 200 docs · DP-Bench · Score 0–1
Nutrient --vision Licensed 0.93
docling 0.89
Nutrient (default) Free 0.89
pymupdf4llm 0.86
opendataloader 0.83
markitdown 0.59
pypdf 0.58
liteparse 0.57

--vision is a licensed tier of the Nutrient document CLI; the CLI's default engine is free for up to 1,000 documents per month (separate from the Data Extraction API free tier above).

Scores are the DP-Bench composite. Per metric, --vision leads outright: 0.96 reading order (NID), 0.94 table structure (TEDS), 0.87 heading hierarchy (MHS) — at 0.35 s/page, faster than docling. Benchmarked 2026-07-06 on an Apple M3 Ultra.

Showcase

Built with Nutrient

See all customer stories
Scaled legal document workflows 50 percent month over month

Replaced a patchwork of open source PDF tooling with Nutrient’s Web SDK and Document Engine — absorbing Harvey’s growth without adding internal infrastructure.

FULL STORY
Scaled to 50,000 students in a single day

Subject scaled 5× with Nutrient’s PDF SDK, handling peak usage without rebuilding its document layer or managing additional infrastructure.

FULL STORY
Powering eSignatures at enterprise scale

DocuSign uses Nutrient to streamline eSignature workflows, enhance performance, and ensure a reliable and customizable document experience for millions of users.

FULL STORY

Use case finder

What do you want to do with your documents?

Extract data

curl -X POST https://api.nutrient.io/build \
-H "Authorization: Bearer your_api_key_here" \
-o result.json \
--fail \
-F invoice=@invoice.webp \
-F instructions='{
"parts": [
{
"file": "invoice"
}
],
"output": {
"type": "json-content",
"keyValuePairs": true
}
}'

Turn PDFs, scans, forms, invoices, contracts, and Office files into typed JSON or LLM-ready Markdown. Preserve tables, key-value pairs, line items, coordinates, and confidence signals so downstream agents can validate before acting.

Explore data extraction

Generate PDFs

curl -X POST https://api.nutrient.io/processor/generate_pdf \
-H "Authorization: Bearer your_api_key_here" \
-o result.pdf \
--fail \
-F html=@index.html

Create PDFs from HTML, templates, Markdown, or structured data. Add headers, footers, dynamic content, and accessibility-ready output for reports, agreements, statements, and customer-facing documents.

Explore PDF generation

Convert documents

curl -X POST https://api.nutrient.io/processor/convert_to_pdf \
-H "Content-Type: application/vnd.openxmlformats-officedocument.wordprocessingml.document" \
-H "Authorization: Bearer your_api_key_here" \
-o result.pdf \
--fail \
--data-binary @document.docx

Convert between PDF, Word, Excel, PowerPoint, HTML, images, and Markdown while preserving layout and structure. Prepare documents for review, storage, automation, or AI ingestion.

Explore conversion

OCR scanned files

curl -X POST https://api.nutrient.io/processor/ocr \
-H "Authorization: Bearer your_api_key_here" \
-o result.pdf \
--fail \
-F file=@document.pdf \
-F data='{
"language": "english"
}'

Make scans, images, and photographed pages searchable, extractable, and ready for automation. Use OCR as the first step in document intake, classification, extraction, and review workflows.

Explore OCR

View documents in my app

curl -X POST https://api.nutrient.io/processor/generate_pdf \
-H "Authorization: Bearer your_api_key_here" \
-o result.pdf \
--fail \
-F html=@index.html

Embed a production-ready document viewer for human review. Let users view, annotate, fill forms, compare, collaborate, and approve documents inside your product.

Explore document viewer

Sign documents

curl -X POST https://api.nutrient.io/sign \
-H "Authorization: Bearer your_api_key_here" \
-o result.pdf \
--fail \
-F file=@document.pdf \
-F data='{
"signatureType": "cades",
"cadesLevel": "b-lt"
}'

Add signing and approval flows to generated or uploaded documents. Support governed completion with audit trails and downstream workflow handoff.

Explore eSignatures

Redact sensitive data

curl -X POST https://api.nutrient.io/processor/redact \
-H "Authorization: Bearer your_api_key_here" \
-o result.pdf \
--fail \
-F file=@document.pdf \
-F data='{
"strategy": "text",
"strategyOptions": {
"text": "macaque",
"caseSensitive": false
},
"redactionState": "apply"
}'

Find and remove sensitive information such as names, emails, SSNs, account numbers, and custom patterns. Build secure, audit-ready workflows for regulated document handling.

Explore redaction

Validate compliance

curl -X POST https://api.nutrient.io/validate_pdfa \
-H "Authorization: Bearer your_api_key_here" \
-o result.json \
--fail \
-F file=@pdfa.pdf

Check PDF/A and PDF/UA requirements, accessibility readiness, archival formats, and compliance-sensitive outputs before documents move forward.

Explore compliance

Open governance

Verify what the AI says — before you act on it

AI can read a document. Knowing whether its answer is actually backed by the source — especially a number in a table — is the hard part. So we open-sourced grounding-en (Apache-2.0), a model that scores whether a claim is entailed by its source document, benchmarked on the open grounding leaderboard.

#1 & #2
Ranked on the open grounding leaderboard (July 2026)
16×
Context length — 8,192 tokens vs the usual 512
0.4B
Parameters — small enough to self-host
0.92–0.97
Number grounding, vs 0.45–0.66 for general NLI


Frequently asked questions

What is Nutrient Document Web Services (DWS)?

Nutrient DWS is a cloud platform of document APIs: Processor for headless document workflows (generation, conversion, OCR, redaction), Data Extraction for structured data from any document, Viewer for embedding document viewing and review in your app, and Accessibility for PDF/UA tagging and validation — all without managing infrastructure.

How does pricing work?

Pricing depends on the API you use. Processor, Viewer, Accessibility, Data Extraction, signing, and enterprise deployments may have different pricing models, limits, and evaluation paths. See the pricing page for current plans, limits, and evaluation paths for each API.

Do I need to manage any infrastructure or servers?

No. Every DWS API is fully hosted and managed — you integrate with a few lines of code and Nutrient handles scaling, security, updates, and maintenance. For teams with data-residency requirements, on-premises and dedicated deployments are available.

Is the API SOC 2 compliant? HIPAA?

Yes — It’s SOC 2 Type 2 audited, HIPAA-eligible, and GDPR-compliant. All API traffic is TLS-encrypted. Hosted in US and EU regions, with on-premises and dedicated deployments available.

Can Nutrient support agentic document workflows?

Yes. Nutrient APIs help agents and applications work with documents in predictable ways: extracting structured data, converting files, generating documents, OCRing scans, redacting sensitive content, validating compliance, and preparing documents for human review. Responses can include structured data, Markdown, coordinates, and confidence signals so downstream systems can check outputs before acting.

What programming languages and frameworks are supported?

The APIs are language-agnostic and work from any stack that can make an HTTP request — Node.js, Python, Java, .NET, PHP, Ruby, Go, and more. Client SDKs and code samples are available in the documentation, and no framework-specific integration is required.

What file formats do Nutrient APIs support?

Nutrient APIs support common document formats, including PDFs, images, Microsoft Office files, HTML, Markdown, and more. Supported formats vary by API, so check the relevant API documentation for exact input and output options.

What kind of support do you provide?

All paid plans include technical support from the Nutrient team, with priority response times for enterprise customers. Every plan gets documentation and integration guides, code examples, and the Discord community.

Build your first document workflow today

Get your API key. Make your first call in 60 seconds. Scale when you’re ready.