Cookie preferences

Essential cookies stay on for login, security, and core site behavior. Analytics cookies are optional and help us improve the product.

Privacy policy

graflows
New: schema memory with approved examples

Turn Your Documents into Structured Data that gets sharper with review.

API-first, layout-aware extraction for the modern stack. Save schemas, approve corrected outputs, and let similar documents reuse your best examples automatically.

invoice_1024.pdf
POST /v1/extract
{
  "document_id": "doc_123xyz",
  "status": "completed",
  "review_status": "pending",
  "memory_source_file_id": "file_98abc",
  "memory_similarity": 0.94,
  "data": {
    "usage": {
      "pages": 1,
      "tokens": 420
    },
    "fields": {
      "invoice_number": "INV-2024-001",
      "date": "2024-02-15",
      "total_amount": {
        "value": 1250.00,
        "currency": "USD"
      },
      "line_items": [
        {
          "description": "Pro Plan Subscription",
          "quantity": 1,
          "amount": 999.00
        },
        {
          "description": "Usage Overage",
          "quantity": 12,
          "amount": 251.00
        }
      ]
    }
  }
}
Memory Match

Reused an approved invoice example from the same saved schema.

Similarity 0.94

Document Memory

Close the loop between extraction and approval.

Graflows can reuse a user-approved file as a one-shot example for future extractions on the same saved schema, improving consistency without changing your API flow.

01

Save the schema

Turn memory on once per schema so every upload knows where reusable examples belong.

02

Approve the result

Review the normalized output and approve the corrected JSON when the extraction is right.

03

Reuse the example

When a similar document appears, Graflows injects the approved result into the prompt as context.

Compliance-First Redaction.

Maintain strict data privacy and compliance by automatically masking sensitive information across your document workflows.

Smart PII Detection

Automatically identify names, emails, addresses, SSNs, and other sensitive entities using advanced NLP models.

Permanent Masking

Go beyond overlays. Our API performs bit-level redaction on PDFs, ensuring the original data cannot be recovered.

document_redacted.pdf
Sensitive Found

Fair, Transparent Pricing.

Scale without surprises. Our usage-based model grows as you grow, managed through a powerful integrated dashboard.

Free + Pay as you go

$0

100 pages included monthly, then $0.05 per page.

What's included

API access
Standard extraction
Community support
Most Popular

Pro

$99/month

5,000 pages included monthly.

What's included

Faster processing
Priority queue access
Email support

Enterprise

Custom

Tailored solutions for high-volume document processing.

What's included

Unlimited volume discounts
SLA-backed performance
Custom model tuning
SSO & Role-based security
Dedicated success manager
On-prem deployment options

Full Control via Dashboard

Track every extraction, manage API keys, and monitor your usage in real-time. Our upcoming dashboard gives you deep visibility into your document workflows.

Real-time Analytics
Key Management
usage_mtd1,248 pgs
credits_left$75.40
statusactive_pro

Common Questions

How is usage calculated?

We count every page processed by our API. A 'page' is defined as a single side of a document (PDF or image). Empty pages are not counted.

What happens if I exceed my quota?

On the Pro plan, you'll be charged $0.05 for every additional page processed. On the Free plan, requests will be capped until the next billing cycle.

Do you offer annual discounts?

Yes! Contact our sales team for custom annual pricing packages with higher volume discounts.

Need more details?View Docs Support