FREQUENTLY ASKED
QUESTIONS
Quick answers to whatever questions you might have.
have questions regarding Api's ? - refer Api docs
DocuMind is an AI-powered Intelligent Document Processing (IDP) platform that automatically extracts, classifies, and analyzes data from any document type. We help businesses eliminate manual data entry, reduce processing time by up to 90%, and achieve 95-99% extraction accuracy. Whether you're processing invoices, contracts, forms, or proprietary documents, DocuMind transforms unstructured documents into clean, structured data in seconds.
Intelligent Document Processing combines AI technologies—Optical Character Recognition (OCR), Natural Language Processing (NLP), and Machine Learning—to automate document workflows. Unlike traditional OCR that simply scans text, IDP understands context, classifies documents, extracts specific fields, validates data, and continuously learns to improve accuracy. It handles structured (forms), semi-structured (invoices), and unstructured (emails, contracts) documents seamlessly.
Key Differentiators:
- • Fastest Training: Train custom models with just 5-10 sample documents (competitors need 100+ samples and weeks of setup)
- • Simple REST API: No complex SDKs to maintain—pure JSON integration that works with any tech stack
- • Rapid Deployment: From signup to first extraction in under 5 minutes
- • SMB-Friendly: Transparent pricing and self-serve platform (not just enterprise-focused)
- • Prebuilt + Custom: Start instantly with prebuilt models OR train custom models for unique formats
- • True Self-Learning: Models auto-improve from every user correction with zero downtime
Prebuilt Models Support: Invoices (single & multi-page, all formats), Receipts (retail, restaurant, travel), Bank statements (all major banks), Payslips/Salary slips, Purchase orders, ID cards and KYC documents.
Custom Models Support: Contracts and legal documents, Medical records and prescriptions, Insurance claims and policies, Shipping and logistics documents, Internal forms and templates, Academic transcripts, Any proprietary document format.
File Formats Accepted: PDF, PNG, JPG, JPEG, TIFF, BMP, GIF (scanned, photographed, or digital)
Our extraction accuracy ranges from 95% to 99% depending on document quality and complexity. Every extracted field includes a confidence score (0–1) so you know exactly which results to trust. Fields with confidence below your threshold (e.g., 80%) are automatically flagged for human review.
- • Standard invoices: 98-99% accuracy
- • Handwritten documents: 90-95% accuracy
- • Complex multi-page documents: 95-97% accuracy
- • Custom trained models: 96-99% after 10 corrections
Processing Speed:
- • Most documents: Under 10 seconds
- • Large documents (50+ pages): 20-30 seconds
- • Batch uploads (100 documents): 5-10 minutes total
No coding required! Our platform is designed for both business users and developers.
For Business Users: Drag-and-drop document upload, Click-to-label custom model training, Visual review interface for validation, One-click export to Excel/CSV/JSON, No programming knowledge needed.
For Developers: Simple REST API with JSON responses, No SDKs to install or maintain, Comprehensive API documentation, Code examples in Python, JavaScript, cURL, Webhook support for automation.
Yes! This is our breakthrough technology. While traditional ML platforms require 100-500 training samples and weeks of work, our proprietary AI needs just 5-10 sample documents and 20 minutes to create production-ready models.
Why This Works: Our AI is purpose-built for document understanding. It already knows how documents are structured (layouts, text patterns, field relationships) from training on millions of real-world documents. When you provide 5-10 samples, you're teaching it YOUR specific format—not teaching it what a document is.
- • Minimum: 5 documents
- • Optimal: 8-10 documents
- • Diversity: Upload samples with variations (different vendors, dates, layouts) for best initial accuracy
Simple 4-Step Process:
- • Step 1: Upload Samples (5 minutes) - Upload 5-10 sample documents of your format
- • Step 2: Define & Label Fields (10 minutes) - Click on fields you want to extract
- • Step 3: Train Model (1 minute) - Click 'Train Model' and system processes in 30-90 seconds
- • Step 4: Deploy & Use - Upload new documents—model extracts automatically
No problem! You have two options:
Option 1: Single Flexible Model - Upload samples representing all format variations. Our AI learns to handle diversity within one model.
Option 2: Multiple Models - Create separate models for distinctly different formats. Unlimited models are supported—each with its own extraction logic.
Continuous Self-Learning: Every time you review and correct an extraction, the system learns from your correction immediately. Future documents benefit from this learning with accuracy improving automatically.
- • Initial training (10 samples): 94% accuracy
- • After 20 corrections: 96% accuracy
- • After 50 corrections: 98% accuracy
- • After 100+ documents: 99%+ accuracy on standard fields
Yes, anytime! You have full control with automatic updates where models auto-learn from corrections with no action needed. Manual retraining allows you to add new sample documents anytime, adjust field definitions, and update validation rules. Each model version is tracked with the ability to rollback if needed.
REST API: Simple POST to upload documents, GET to retrieve results. Works with any programming language or platform. Full documentation available at docs.DocuMind.com/api with code examples provided in Python, JavaScript, cURL, and more.
Yes! Most modern systems (Salesforce, SAP, QuickBooks, Xero, NetSuite) support REST APIs. Your developers can connect DocuMind directly.
For large-scale integrations, our Enterprise Support team provides integration consultation, custom connector development, technical support during implementation, and testing assistance. Contact enterprise@DocuMind.com for dedicated integration support.
No! You only pay for successful extractions.
What's Free: Failed extractions (poor image quality, unsupported format), Re-processing the same document after corrections, Test uploads and sample processing, API status checks and document queries.
Enterprise-Grade Security:
- • Encryption: TLS 1.3 encryption in transit, AES-256 encryption at rest, End-to-end encryption for sensitive documents
- • Access Controls: Role-Based Access Control (RBAC), Multi-factor authentication (2FA), IP whitelisting (Enterprise), API key rotation
- • Infrastructure: Hosted on Microsoft Azure (ISO 27001, SOC 2 Type II certified), Data centers in India, US, EU
- • Compliance: SOC 2 Type II, GDPR compliant, HIPAA ready, ISO 27001 certified
Automatic Deletion Policy: Documents retained for 7 days after processing (for quality review and troubleshooting). After 7 days, they are automatically deleted from all systems. Backups are purged after 30 days.
Your Control: Set auto-delete timelines (1 day, 7 days, 30 days), Bulk delete old documents, Request immediate deletion via support.
What We DON'T Do: Never use your documents for model training without explicit consent, Never share documents with third parties, Never sell or monetize your data.
Yes! Our OCR technology processes handwritten forms (cursive or print), mixed handwritten + printed documents, signatures (for verification), and scanned notes and annotations.
- • Printed text: 98-99% accuracy
- • Clear handwriting: 90-95% accuracy
- • Complex/messy handwriting: 85-90% accuracy
Yes! We support 50+ languages including:
Indian Languages: Hindi, Tamil, Telugu, Marathi, Bengali, Gujarati, Kannada, Malayalam, Punjabi, Urdu
Global Languages: Spanish, French, German, Italian, Portuguese, Russian, Chinese, Japanese, Korean, Arabic, and more.
Our AI handles documents with multiple languages (e.g., English + Hindi invoice). Custom models can be trained on documents in any language.
Built-In Correction System: Split-screen view with original document and extracted data. Click any field to correct with the system highlighting low-confidence fields automatically. Every correction teaches the model and future documents benefit from your feedback with no manual retraining needed.
Validation Rules: Set business rules (e.g., 'Invoice total must equal line items sum'), Auto-flag violations for review, Cross-field validation.
Yes! Batch processing is fully supported with drag-and-drop (up to 1,000 files at once), Cloud folder sync (auto-process new files), Email forwarding (send to process@DocuMind.com), and API batch upload endpoint.
- • 100 documents: ~10 minutes
- • 1,000 documents: ~1 hour
- • 10,000+ documents: Contact us for dedicated processing
Yes, multiple export formats are supported:
Supported Formats: Excel (.xlsx), CSV, JSON, XML for legacy system integration.
Export Options: Single document export, Batch export (all documents in folder), Scheduled exports (daily/weekly), API-based real-time export, Webhook push to your system.
Customization: Choose which fields to export, Define column order, Apply filters (date range, document type), Merge tables across pages.
