scanix
Download
v2.0 — out now

Document capture, OCR, and AI extraction that runs on your machine.

Scanix Desktop scans paper and PDFs, extracts structured data with on-device AI, and never sends your documents to a third party. Built for compliance teams and regulated industries that treat their data as their own.

Card required · No charge until trial ends · Cancel anytime

scanix-desktop · INVOICE_2026-Q1.pdf
Acme Corp
Invoice #INV-2026-0142
Total: $12,400.00
Extracted fields
VendorAcme Corp99%
Invoice #INV-2026-014298%
Total$12,400.0097%
Due date2026-06-0195%
All fields validated locally
GDPR by architectureHIPAA-compatible deploymentsSOC 2 controlsAir-gapped capablePDF/A-3 archival

Three things we refuse to compromise on

Scanix is built around the boring guarantees compliance teams actually ask for.

Your data, your machine

Local AI is a first-class option, not an afterthought. Choose between BYO cloud keys or fully on-device inference — the data path is yours to decide.

Production-grade OCR

Tesseract, PaddleOCR, and curated vision models bundled. TWAIN/WIA scanners, hot folders, and 9 built-in document templates.

Built for audit

Page hashing at ingest, JSON Lines audit logs, hOCR / ALTO / PDF/A-3 export. The compliance trail is automatic, not bolted on.

For teams who can't send documents to anybody

Scanix runs entirely behind your firewall. Three of the industries we hear from most.

Talk to us

Healthcare

Patient intake forms, lab reports, claim documents. PHI never leaves your network.

Legal & Compliance

Contract digitization, KYC packets, audit trails with page-level hashing.

Financial Services

Statements, IDs, vendor invoices. On-device classification across 9 templates out of the box.

Templates that fit your workflow

JSON in, JSON out. No black box.

Define your document classes in a JSON template; Scanix routes incoming pages through OCR, classification, and field extraction. Every step is observable, versioned, and reproducible.

Read the docs
{
  "name": "vendor_invoice",
  "fields": [
    { "key": "vendor",     "type": "string",  "required": true },
    { "key": "invoice_no", "type": "string",  "regex": "^INV-" },
    { "key": "total",      "type": "money",   "currency": "USD" },
    { "key": "due_date",   "type": "date",    "format": "iso8601" }
  ],
  "extractor": "scanix:vision-7b",
  "post_processors": ["normalize_dates", "validate_currency"]
}

How it works

01

Install in under a minute

Single-file installer for Windows. Signed, notarized, no admin rights for the user-mode portion.

02

Point at your sources

TWAIN scanners, WIA devices, hot folders, drag-and-drop. PDFs, images, multi-page TIFFs all welcome.

03

Export structured data

JSON, CSV, hOCR/ALTO XML, or PDF/A-3 with embedded extraction. Your downstream pipeline stays unchanged.

Pricing that scales with seats, not surprises

Per-device seats, monthly or yearly. Annual saves about 20%. Perpetual licenses available for Standard and Pro on request.

Try Scanix on your own documents

15 days of full access. We won't charge your card until the trial ends — cancel anytime. Your data never leaves your machine.