On-Device Field Extraction by Verify

Product Introduction

1. Definition: On-Device Field Detection is a new capability within the Veryfi Lens SDK, a mobile and browser-based camera SDK for document capture. It is a real-time, on-device optical character recognition (OCR) and data validation engine that processes documents locally on a user's smartphone or browser.
2. Core Value Proposition: It exists to eliminate the latency and inefficiency of cloud-based document processing by validating critical data fields (like total amount, date, merchant) at the precise moment of capture, before upload. This ensures data quality at the source, reduces backend processing errors, and creates a superior, instantaneous user experience for mobile data capture workflows.

Main Features

1. Real-Time On-Device Validation: The SDK performs OCR and field extraction directly on the user's device using optimized, in-house AI models. It validates mandatory fields against predefined rules (e.g., is a total amount present and legible?) and provides immediate visual feedback to the user, such as highlighting missing or unreadable fields. This happens without requiring an internet connection for the validation step.
2. Seamless Integration with Veryfi AI Cloud: After on-device validation passes, the document can be uploaded to Veryfi's cloud APIs for full, high-accuracy data extraction, including line-item details, tax calculations, and complex field parsing. This hybrid approach combines the speed of local processing with the power of cloud-based AI.
3. White-Label & Customizable SDK: The Veryfi Lens SDK is provided as a white-label solution, allowing developers to fully customize the camera UI, validation behaviors, and user interface to match their brand. It supports advanced camera controls like auto-capture, edge detection, perspective correction, and flash control.
4. Multi-Document Type Support: While prominently featured for receipts, the underlying technology and SDK are designed for a wide array of documents. The platform's APIs support extraction from invoices, purchase orders, bank statements, checks, W-2s, insurance cards, and many more, making the on-device capture a universal entry point.

Problems Solved

1. Pain Point: High Rejection Rates in Post-Upload Processing. Traditional mobile capture sends all images to the cloud first. Poor-quality images or missing fields are only discovered after upload, leading to high rejection rates, user frustration, and backend processing waste.
2. Target Audience: Mobile App Developers (FinTech, Expense Management, Accounting SaaS), Product Managers for B2B SaaS platforms, and Enterprise Architects building document-intensive workflows in industries like insurance (claims), banking (remote deposit capture), healthcare (patient intake), and construction (field reports).
3. Use Cases:
   • Expense Management Apps: Ensuring a receipt photo contains a legible total and date before allowing submission, guaranteeing clean data for reimbursement.
   • Insurance Claims Processing: Validating that a photo of a damage report or bill includes a policy number and claim date on-site, accelerating adjuster workflows.
   • Banking & KYC: Capturing a government ID or bank statement and immediately confirming all required fields are present for compliance, reducing onboarding drop-off.
   • Accounts Payable Automation: Guaranteeing an invoice captured by a field agent has a vendor name and invoice number before syncing to the central AP system.

Unique Advantages

1. Differentiation: Unlike pure cloud OCR APIs (e.g., Google Vision, Amazon Textract) which only process after upload, and unlike simple camera SDKs that only capture images, Veryfi Lens with On-Device Field Detection adds an intelligent validation layer before the cloud round-trip. It also differs from other mobile scanning SDKs by being deeply integrated with a specialized, high-accuracy document-specific AI backend.
2. Key Innovation: The hybrid on-device/cloud architecture. By running a lightweight but capable validation model directly on the device, it solves the "garbage in, garbage out" problem at the source. This "validate first, process later" model is a significant shift from industry norms and drastically improves overall system efficiency and user satisfaction. The use of in-house foundation AI models (not generic OCR) trained specifically on financial and business documents is another key advantage for accuracy.

Frequently Asked Questions (FAQ)

1. How does on-device field extraction work without an internet connection? The Veryfi Lens SDK packages a compact, optimized AI model directly into your mobile app or web application. This model runs locally on the device's processor (CPU/GPU), analyzing the camera feed in real-time to detect document boundaries, perform initial OCR, and check for required data fields, all without needing network connectivity.
2. What is the difference between Veryfi Lens SDK and the Veryfi API? The Veryfi Lens SDK is a front-end capture library you embed in your mobile or web app to acquire and validate documents. The Veryfi API is a back-end, cloud-based service that performs deep, structured data extraction from the validated images sent by the SDK. They are complementary products designed to work together for an end-to-end solution.
3. Is the data processed on-device secure and private? Yes. Since the initial validation and field detection occur locally, the sensitive document image and data never leave the user's device until they explicitly choose to upload it. This aligns with data sovereignty requirements and reduces privacy surface area. Veryfi is also SOC 2 Type II certified and compliant with GDPR, HIPAA, and CCPA.
4. Can I customize what fields are validated on-device? Yes. The Veryfi Lens SDK provides developers with configuration options to define which fields (e.g., total, date, merchant, tax) are mandatory for validation. You can tailor the logic to match your specific application's data requirements.
5. Does this work for all document types, or just receipts? While the on-device validation is highly effective for receipts, the SDK is a general-purpose document capture tool. The validation logic can be adapted, and the captured image is then sent to the appropriate Veryfi API endpoint (e.g., for invoices, checks, IDs) for full, type-specific data extraction in the cloud.