Veryfi
by Veryfi · Invoice & Document OCR
AI-powered OCR API and SDKs that turn documents into structured data
- Works with
- QuickBooks Online, QuickBooks Desktop, Xero, NetSuite, SAP, Zapier, n8n
- Deployment
- Cloud
- Company size
- SMB, Mid-market, Enterprise
- Pricing
- Tiered subscription with per-document and volume-based pricing
- Founded
- 2017
- Headquarters
- San Mateo, California, United States
Overview
Veryfi is a document data extraction platform built around an OCR API and a set of mobile and browser document-capture SDKs. It converts unstructured documents — receipts, invoices, bank checks, bank statements, purchase orders, W-2s, W-9s, and dozens of other types — into clean, structured JSON, extracting fields such as vendor details, line items, totals, dates, and tax amounts. The platform is aimed primarily at developers and product teams who want to embed document intelligence into their own applications rather than buy a packaged AP or expense product.
The core of the offering is a RESTful API paired with native SDKs across mobile (Swift, Kotlin, Objective-C), cross-platform frameworks (React Native, Flutter, Xamarin, Ionic, Cordova), and backend languages (Python, Node.js, PHP, Ruby). Veryfi positions its models as pre-trained, so customers can begin extracting data without building or labeling a training set, while a model-training API lets teams submit corrections as feedback. Beyond raw extraction, Veryfi has added a layer of data-extraction "agents" for tasks such as document classification, PDF splitting, fraud and GenAI detection, and document-to-markdown conversion.
Veryfi serves accounting, fintech, banking, construction, healthcare, real estate, and CPG use cases, with common applications in accounts-payable automation, expense management, remote deposit capture, KYC/KYB, and loyalty programs. Documents flow into a Hub command center after being captured via the Lens SDK, email, or API, and can be routed into downstream accounting and ERP systems either through prebuilt connectors (QuickBooks, Xero) or by mapping the JSON output to a target system's API.
Screenshots & demo
Demo video from the vendor's YouTube channel. Screenshots sourced from Veryfi.
Features & capabilities
Data extraction API
Core OCR and document-to-data processing.
- Structured JSON output with vendor, line items, totals, dates, and tax fields
- Line-item level extraction from receipts and invoices
- Pre-trained models requiring no initial training data
- 40+ supported document types out of the box
- Custom field extraction
- Sub-second to seconds-level processing claims
- Duplicate-document detection flag
Document capture SDKs (Lens)
Embeddable camera capture for mobile and browser.
- Lens for Mobile (iOS and Android) camera capture SDK
- Lens for Browser web capture
- On-device machine-learning model for framing and quality checks
- Capture in varied lighting and on any surface
- Native SDKs for Swift, Kotlin, and Objective-C
- Cross-platform support via React Native, Flutter, Xamarin, Ionic, and Cordova
Supported document types
- Receipts and invoices
- Bank checks and bank statements
- Purchase orders and bills of lading
- W-2s, W-9s, and W-8BEN-E tax forms
- Credit cards and business cards
- Hotel folios and healthcare insurance cards
AI agents and intelligence
Higher-level processing built on the extraction engine.
- Document classification
- Fraud detection suite
- GenAI / synthetic-document detector
- PDF splitter for multi-document files
- Document-to-markdown conversion
- Product intelligence (SKU/UPC expansion)
Platform and developer tooling
- Hub command center with document inbox and dashboard
- RESTful API with SDKs across 12+ languages
- Model-training API for submitting corrections as feedback
- Multi-channel ingestion via Lens, email, and API
- No-code workflow and embedded options
- API Hub access on paid plans
Common use cases
- Accounts payable invoice automation
- Employee expense management and receipt capture
- Remote deposit capture for bank checks
- KYC/KYB identity and document verification
- Loyalty and rewards programs from receipt scanning
- Insurance claims document processing
- Embedding document capture into a SaaS or mobile app
Strengths & considerations
Strengths
- Developer-first API + embeddable capture SDK rather than a packaged end-user app
- Pre-trained models that work without customer-supplied training data
- Native SDK coverage across mobile and cross-platform frameworks
- Line-item extraction and duplicate detection built into the API
- Self-hosted GPU infrastructure with a no-humans-in-the-loop processing claim
Considerations
- Paid plans start at a $500/month minimum, which reviewers note is steep for low-volume or small-business use
- Multi-language support is cited by some reviewers as an area to improve
- Geared toward teams with development resources; no-code paths are more limited than the API
- Prebuilt ERP connectors are limited (mainly QuickBooks and Xero); NetSuite/SAP typically require custom API mapping
- Some reviewers report support quality has declined
ERP integrations
Real-time receipt and invoice sync into QuickBooks Online.
Receipt collection and expense data sync to QuickBooks Desktop.
Captured documents sent automatically to a Xero account.
Approved/extracted data mapped to NetSuite via Veryfi's JSON output and the ERP's API.
Structured JSON mapped into SAP through API integration; no named prebuilt connector.
Verified Veryfi integration on the n8n workflow automation platform.
Pricing
Free tier covers up to 100 documents/month. Starter (~$500+/month minimum) includes roughly 5,000 documents with published per-document rates (e.g., invoices ~$0.16, receipts ~$0.08, bank checks/statements ~$0.25). Growth tier is volume-based with SSO, SLAs, custom retention, and model training; contact sales. 14-day free trial, no credit card required. Get an independent shortlist with pricing guidance below.
Technical & security
- Hosting
- Vendor-hosted on proprietary GPU infrastructure (NVIDIA DGX H100)
- Compliance
- SOC 2 Type II, GDPR, CCPA, HIPAA
- Mobile app
- Yes
- Languages
- English
About the vendor
- Founded
- 2017
- Headquarters
- San Mateo, California, United States
- Employees
- ~50
- Ownership
- Private (venture-backed; investors include NewView Capital, Act One Ventures, and Y Combinator)
- Notable customers
- Navan, Rippling, Volvo
Alternatives to Veryfi in Invoice & Document OCR
Veryfi — frequently asked questions
What documents can Veryfi extract data from?
Veryfi supports 40+ document types including receipts, invoices, bank checks, bank statements, purchase orders, bills of lading, W-2s, W-9s, W-8BEN-E forms, credit cards, business cards, hotel folios, and healthcare insurance cards. Each returns clean, structured JSON with fields such as vendor, line items, totals, dates, and taxes.
Does Veryfi connect to ERP and accounting systems?
Veryfi offers prebuilt connectors for QuickBooks Online, QuickBooks Desktop, and Xero. For NetSuite, SAP, and other systems, its structured JSON output is mapped to the target system's API, and it also integrates with Zapier and n8n for workflow automation.
How is Veryfi priced?
Veryfi has a free tier for up to 100 documents per month. Paid plans start at a Starter tier around $500+/month (roughly 5,000 documents) with published per-document rates, and a volume-based Growth tier with SSO, SLAs, and model training. A 14-day free trial is available without a credit card.
Is Veryfi secure and compliant?
Veryfi is SOC 2 Type II certified and states GDPR, CCPA, and HIPAA compliance (it will sign a BAA on request). It encrypts data in transit (TLS 1.2/1.3) and at rest (AES), and processes documents on self-hosted GPU infrastructure with no humans in the loop.
Do I need to train Veryfi's models?
No. Veryfi's models are pre-trained, so extraction works from the moment you integrate. A model-training API is available to submit corrections as feedback and improve accuracy for specific document types over time.
Evaluating Invoice & Document OCR?
Tell us your ERP and requirements and we'll send an independent shortlist — including Veryfi and the best-fit alternatives — with honest pros and cons.