Document Intelligence MCP
Document OCR, classification, table extraction, and summarization using local AI vision. Supports invoices, contracts, forms, and reports in PDF, PNG, TIFF, and JPEG. Returns structured JSON — no cloud upload required.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
document_url |
string | ✅ | URL or base64 of document image/PDF |
tasks |
array | ✅ | Tasks: ocr, classify, extract_tables, summarize |
doc_type_hint |
string | ❌ | Hint: invoice, contract, form, report |
language |
string | ❌ | Document language ISO code (auto-detected) |
Example Response
{
"classification": "invoice",
"confidence": 0.98,
"ocr_text": "INVOICE #INV-2024-0847
Date: March 15, 2024
Vendor: Acme Supplies Ltd.",
"tables": [
{
"headers": ["Item", "Qty", "Unit Price", "Total"],
"rows": [["Widget A", "100", "$4.50", "$450.00"], ["Shipping", "1", "$25.00", "$25.00"]]
}
],
"summary": "Invoice from Acme Supplies for 100 Widget A units plus shipping. Total: $475.00"
}
Use Cases
- Accounts payable invoice processing automation
- Legal contract review and clause extraction
- Medical records digitization for EMR systems
Access
# x402 endpoint — pay $0.05 USDC per call (Base mainnet)
POST https://x402.ntriq.co.kr/document-intel
# Service catalog
curl https://x402.ntriq.co.kr/services
x402 micropayments — USDC on Base, gasless EIP-3009