DOCUPIPE
Solutions
Resources
Pricing
Document API for Developers
Documents in.
JSON out.
Parse any document. Extract any JSON schema.
One REST API. PDFs, images, Word, Excel, handwriting — anything.
Request
curl -X POST https://app.docupipe.ai/document \
-H "X-API-Key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"document": {
"file": {"url": "https://example.com/invoice.pdf"},
"fileExtension": "pdf"
},
"workflowId": "YOUR_WORKFLOW_ID"
}'200 OK
Webhook payload
{
"eventType": "standardization.processed.success",
"documentId": "dcmt5678",
"data": {
"invoiceNumber": "08472",
"vendor": "Martinez & Sons",
"date": "2025-03-14",
"total": 907.76,
"currency": "USD"
}
}One POST with a workflow ID. Structured JSON delivered to your webhook when ready.
99%
Extraction Accuracy
1000+
Teams Using DocuPipe
<5s
Per Page Processing
Over 1 billion pages processed, and counting
Trusted by customers big and small across every industry
Rated 4.9/5 on G2 verified reviews
The Problem
Most document APIs stop at OCR
OCR is the easy part of document AI. The hard part is everything that comes after — and most APIs leave it to you. You paid for a document API and got an OCR engine plus a long list of TODOs.
Most APIs return raw text
You get a wall of unstructured text and have to write your own field extraction. Schema-aware extraction is bolted on as a separate service if it exists at all.
Multi-step pipelines
Upload to one service, OCR through another, run forms-analysis through a third, glue it together yourself. Each hop adds latency and a new failure mode.
Handwriting and scans break
Most APIs handle clean digital PDFs and fall over on real-world inputs: scanned forms, handwritten notes, photographed documents, faxed pages.
Compliance is an afterthought
HIPAA, SOC 2, BAAs, zero data retention — table stakes for production document workloads, often missing or gated behind enterprise sales motions.
Built for every document
You got it, we parse it
One API across every doc type. No specialized models to wire up, no per-vertical pipelines to maintain.
Invoices
Bills of lading
Medical records
Bank statements
Contracts
Leases
Purchase orders
Tax forms
IDs and passports
Lab reports
Insurance claims
Anything else
Capabilities
Parse and extract, one API
A complete set of primitives. Production-ready from day one — no assembly required.
Parse any document
PDFs, images, Word, Excel, scans, handwriting. Clean structured text with tables, layouts, and checkboxes preserved.
Extract any schema
Define a JSON schema once, run it across thousands of documents. Get exact-fit structured data — no post-processing, no glue code.
Tables and layouts
Tables come back as structured JSON arrays. Multi-column layouts parsed in correct reading order. Checkboxes detected.
60+ languages
English, Spanish, French, German, Chinese, Japanese, Korean, Arabic, Hebrew, and more. Multilingual documents handled natively.
One REST API
Single endpoint, async by default, webhooks for completion, base64 or URL upload. Works with any language that speaks HTTP.
Fast and compliant
Under 5 seconds per page on most documents. SOC 2 Type II, ISO 27001, HIPAA. BAAs available for healthcare workloads.
Developer Experience
One POST. Webhook delivers.
Build a workflow once — parse, extract, and any post-processing — then POST documents to it. The structured JSON arrives at your webhook when processing finishes. No polling, no glue code.
extract.py
import requests
# one call. the workflow handles parsing + schema extraction.
# structured JSON is delivered to your webhook when ready.
requests.post(
"https://app.docupipe.ai/document",
headers={"X-API-Key": "YOUR_API_KEY"},
json={
"document": {
"file": {"url": "https://example.com/invoice.pdf"},
"fileExtension": "pdf",
},
"workflowId": "YOUR_WORKFLOW_ID",
},
)
# your webhook receives the structured JSON when processing finishes
@app.post("/webhook/docupipe")
def webhook(payload: dict):
if payload["eventType"] == "standardization.processed.success":
print(payload["data"]) # exact-fit JSON matching your schemaComparison
DocuPipe vs the cloud doc APIs
| DocuPipe | AWS Textract | Google Document AI | Azure Doc Intelligence | |
|---|---|---|---|---|
| Schema-based extraction | Built-in | No (separate Bedrock setup) | Form parser only | Custom model training |
| Handwriting recognition | Yes, production-grade | Basic | Limited | Limited |
| Table extraction | Full structure as JSON | Basic JSON | Basic JSON | Basic JSON |
| File types | PDF, images, Word, Excel | PDF, images | PDF, images | PDF, images |
| Language support | 60+ languages | ~30 languages | ~50 languages | ~70 languages |
| API design | Single REST endpoint | Multiple services | Multiple services | Multiple services |
| Compliance | SOC 2, ISO 27001, HIPAA | SOC 2, HIPAA | SOC 2, HIPAA | SOC 2, HIPAA |
| Pricing transparency | Per-page, on site | Per-feature, complex | Per-feature, complex | Per-feature, complex |
| Free tier | 300 pages, no card | Limited free tier | Limited free tier | Limited free tier |
Pricing
300 free credits. Plans from $99/mo.
No credit card required to start. Per-page transparent pricing — no per-feature surprises.
See PricingFAQ
Frequently asked questions
DocuPipe is a document API for developers. Two core capabilities: parse any document into clean structured text, and extract any JSON schema you define from any document. One REST endpoint, async-by-default with webhooks, works with PDFs, images, Word, Excel, scans, and handwriting.
PDFs (native and scanned), images (PNG, JPG, TIFF, WebP), Word documents (DOC, DOCX), Excel spreadsheets (XLS, XLSX, CSV), plain text, JSON, XML, and HTML. AI-powered OCR handles handwritten content, checkboxes, and low-quality scans.
60+ languages including English, Spanish, French, German, Portuguese, Italian, Dutch, Swedish, Chinese, Japanese, Korean, Arabic, Hebrew, Hindi, Thai, and more. Multilingual documents handled natively.
You define a JSON schema for the structured data you want to extract — fields, types, descriptions, nested arrays, the whole thing. DocuPipe runs it against any document and returns exact-fit JSON matching your schema. You can also generate schemas interactively in the dashboard from example documents.
Most documents process in under 5 seconds per page. Calls are async by default — you get a job ID immediately and either poll for completion or receive a webhook callback when ready. Suitable for both interactive and batch workloads.
Production-grade OCR handles handwritten content, checkboxes, photographed documents, and low-quality scans. Tables and multi-column layouts are preserved in the correct reading order.
SOC 2 Type II, ISO 27001, and HIPAA compliant. BAA agreements available for healthcare workloads. All documents encrypted in transit (TLS) and at rest. Zero-data-retention available. Customer data is never used for model training.
Credit-based: most operations cost 1-2 credits per page. Start with 300 free credits — no credit card required. Paid plans start at $99/mo for 2,500 credits, with volume discounts on higher tiers. Full pricing on the pricing page.
REST API works with any language that speaks HTTP — Python, Node.js, Go, Java, Ruby, PHP, .NET. No-code integrations via Make.com, Zapier, and n8n. Output as JSON, CSV, or Excel — feed directly into your database, vector store, or downstream pipeline.
Yes — sign up gets you 300 free credits and a dashboard where you can upload your own documents and try extraction interactively. No credit card required.
Start building
300 free credits. One REST API. Your documents, your schema, structured JSON back.