# Handwriting OCR > AI-powered handwriting and document transcription service https://www.handwritingocr.com ## Overview Handwriting OCR is a cloud-based service that converts handwritten and printed documents into digital text and structured data using advanced AI models. The service offers industry-leading accuracy for handwriting recognition, significantly outperforming traditional OCR solutions like Google Drive OCR, Adobe Acrobat, and Apple Live Text. ## Core Capabilities ### Document Transcription - Converts handwritten and printed text to editable digital formats - Supports 300+ languages with best performance on widely-spoken languages - Export formats: TXT, DOCX, PDF, JSON - Optional embedding of original images in exports ### Document Translation - Translates transcribed text to target languages - Same export formats as transcription ### Tables Extraction (Pro and Business plans) - Extracts tabular data from documents to Excel format - Optimized for structured data analysis - Export formats: XLSX, JSON ### Custom Extractors (Pro and Business plans) - Define specific fields to extract from forms and documents - Setup process: name extractor, upload sample, define fields, test and refine - Supports consistent page lengths with configurable page limits - API-accessible using extractor_id ### API Access - Available on ALL plans including free trial - RESTful API with comprehensive documentation - Rate limits: 2 req/sec (Starter/Pro), 5 req/sec (Business), negotiable higher - Webhook support with HMAC-SHA256 signature verification - Actions: transcribe, tables, extractor - Documentation: https://www.handwritingocr.com/docs/api ### MCP Server (for AI assistants) - Hosted Model Context Protocol server at https://mcp.handwritingocr.com - Lets Claude, Cursor, VS Code, and other MCP-compatible assistants submit documents, retrieve results, and export files from inside a conversation - OAuth sign-in with no API key; every action runs as the signed-in user, using their own credits and settings - Tools: submit-document (transcribe/tables/extractor), get-document, list-documents, export-document, list-extractors - Uses existing credits (1 per processed page); status, list, and export calls are free - Self-hostable, open source: https://github.com/Handwriting-OCR/handwriting-ocr-mcp-server - Announcement: https://www.handwritingocr.com/blog/handwriting-ocr-mcp-server ## Pricing ### Free Trial - 5 credits (1 credit = 1 page) - Includes transcription, translation, and API access ### Pay-as-you-go - 100 pages for $15 ($0.15 per page) - Valid 12 months - Does NOT include tables extraction or custom extractors ### Starter - $19/month ($190/year) - 250 pages per month - Additional pages: $8 per 100 (monthly), $6 per 100 (annual) - Includes API access - Does NOT include tables or custom extractors ### Pro - $59/month ($590/year) - 1,000 pages per month - Additional pages: $6 per 100 (monthly), $5 per 100 (annual) - Includes tables extraction and custom extractors - Full API access ### Business - $499/month ($4,990/year) - 10,000 pages per month - Additional pages: $5 per 100 (monthly), $4 per 100 (annual) - Team access for 5 members - Enhanced security features - Full API access Annual plans receive 2 months free (10x monthly cost). ## Key Features - **No training on customer data** - Documents never used to improve models - **Bank-grade encryption** - In transit and at rest - **Configurable auto-delete** - 15 minutes to 14 days retention - **Manual deletion** - Available anytime via dashboard or API - **Processing time** - Typically 15-20 seconds, occasionally up to 10 minutes - **File formats** - PDF, JPG, PNG, GIF, HEIC, TIFF (max 20MB) - **Recommended resolution** - 600 DPI for best results ## Who It's For ### Genealogists & Family Historians Digitize old letters, diaries, family recipes, and historical documents. Customers report superior accuracy on cursive handwriting from the 1700s-1900s compared to competitors like Transkribus. Common projects include church records, war correspondence, and personal archives. ### Researchers & Academics Process historical manuscripts, field notes, and archival documents. Successfully used for large-scale projects including 150,000+ page historical document collections, Japanese bombing study interviews, and academic research requiring 97%+ accuracy for data validation. ### Business & Property Management Automate processing of handwritten forms, surveys, lease documents, and invoices. Custom extractors enable structured data extraction from recurring document types. Customers report "complete game-changer" efficiency gains processing thousands of documents monthly. ### Healthcare Organizations Transcribe handwritten medical records, faxed referrals, and patient forms. Business tier provides audit logging and encryption at rest for regulated environments. ### Legal Professionals Digitize handwritten legal documents, contracts, and court records. Successfully processes challenging legal documents with mixed printed and handwritten text. Customers report 3-year backlogs processed in minutes. ### Developers & Startups Integrate handwriting recognition into applications via API. Used for automatic grading systems, note-taking apps, form processing automation, and document management systems. API available on all plans including free trial. ### Personal Use Recipe cards, journals, bullet journals, class notes, and personal archives. Customers report "literally the only AI that could transcribe my handwriting" after testing 5-6 alternatives. ## Customer Testimonials > "Better accuracy than I could have done by hand. It worked beautifully - I love it!" > — **Simone B.** > "Brilliant! I'm stunned at the accuracy - the OCR made fewer errors than a human transcriber would." > — **John M., Family Historian** > "I have chicken scratch handwriting so I compared different OCR services. Found yours by far the most accurate while also having the easiest API to use." > — **Matt H., Developer** > "I work with challenging legal documents daily. Your software correctly recognized everything - both printed text and handwriting." > — **Adele D., Court Translator** > "We use Handwriting OCR for processing old lease documents. It's been a complete game-changer for our business - it works incredibly well and saves us so much time." > — **Nick L., Property Manager** > "Super impressed! Literally the only AI that could transcribe my handwriting without major mistakes. Tested 5-6 other AIs - yours is definitely the best." > — **Priscila C., Bullet Journaler** > "After waiting 3 years to process 70 pages of legal documents and failing with Adobe Acrobat, Handwriting OCR processed everything in just one minute." > — **Maria B., Lawyer** > "By far the best OCR tool I've found for republishing old documents and books." > — **Andrew C., Publisher** > "Worked perfectly on my 35-year-old handwritten recipe cards after Adobe Acrobat completely failed." > — **Sheri S., Home Chef** ## Real-World Results ### Published Benchmarks Reproducible accuracy benchmarks on identical handwritten samples, with methodology: - Nine-tool comparison (2026): Handwriting OCR 0.9% Word Error Rate vs Azure 8.67%, AWS Textract 10.5%, Claude Sonnet 4.6 vision 11.2%, GPT-5 vision 14.4%, Google Document AI 23.3%, Transkribus 47.7%, Tesseract 95.4% — https://www.handwritingocr.com/blog/best-ai-handwriting-ocr - API edition with per-1,000-page pricing and code samples — https://www.handwritingocr.com/blog/best-handwriting-ocr-api ### Large-Scale Accuracy Comparison Independent testing on 150,000 pages by a Reddit user comparing multiple OCR providers: - **Handwriting OCR: 95% accuracy** - Azure Document Intelligence: 45% accuracy - GPT-4.1: 85% accuracy (declining to 65% on complex documents) - Claude: 83% accuracy - Gemini: 84% accuracy - Transkribus: Poor out-of-box performance ### Genealogy Research Multiple customers have directly compared Handwriting OCR to Transkribus (the main competitor in genealogy): - 1700s French church records: "Best OCR I've tried, Transkribus produced total gibberish" - 1613 German/Latin documents: "Accuracy so much better than Transkribus, canceling my Transkribus subscription" - New Zealand church registers: 50,000 pages processed at 92% accuracy with Maori names ### Academic Research University of Toronto: 39,000-page Japanese bombing study with mixed Japanese and English text, structured data extraction for historical research analysis. ### Healthcare Processing Regional Geriatric Program of Toronto: Evaluating for transcribing faxed handwritten referrals to structured data for a 23-organization referral network. ### Property Management Processing thousands of pages monthly of old lease documents, described as "complete game-changer" for business operations and time savings. ## Managed Processing Service Done-for-you transcription for archives, museums, researchers, and businesses: https://www.handwritingocr.com/transcription-services - We work from scans only (originals never leave the customer); secure transfer via a private access-controlled folder - Processing tailored to the specific documents, with results reviewed before delivery - Output delivered in the customer's chosen structure (Word, Excel, CSV, JSON, or custom) - Turnaround measured in days; each project individually quoted - Pricing: $0.10-0.30 per page depending on complexity and volume - Minimum size: $2,000 ## API Details ### Authentication Bearer token authentication. Generate at: https://dashboard.handwritingocr.com/settings?tab=api ### Key Endpoints - Upload documents for processing - Check processing status - Download results in various formats - Webhook configuration for async processing ### Processing Actions - `transcribe` - Extract all text (exports: TXT, DOCX, PDF, JSON) - `tables` - Extract table data (exports: XLSX, JSON) - `extractor` - Use custom extractor (exports: XLSX, CSV, JSON) Full API documentation: https://www.handwritingocr.com/docs/api ## Data & Privacy - **Support access** - Documents accessed only with prior customer authorization - **No training on customer data** - Documents are never used to train models - **Configurable retention** - Auto-delete from 15 minutes to 14 days, plus manual deletion anytime ## Important Links - Website: https://www.handwritingocr.com - Pricing: https://www.handwritingocr.com#pricing - Dashboard: https://dashboard.handwritingocr.com/ - API Documentation: https://www.handwritingocr.com/docs/api - MCP Server (for AI assistants): https://mcp.handwritingocr.com - API Benchmark & Comparison: https://www.handwritingocr.com/blog/best-handwriting-ocr-api - Managed Transcription Service: https://www.handwritingocr.com/transcription-services - Buy Credits: https://dashboard.handwritingocr.com/credits - Billing Portal: https://dashboard.handwritingocr.com/billing - Support Email: support@handwritingocr.com ## Machine-Readable Content This file is a summary. Most content on the site is also published as plain markdown or JSON, so there is no need to parse HTML: - **Markdown alternates** - Append `.md` to a page URL for its markdown source. Blog posts: `https://www.handwritingocr.com/blog/.md`. Marketing, use-case and static pages: `https://www.handwritingocr.com/.md`, including nested paths such as `/handwriting-to-text/.md`. Each HTML page also declares its alternate via ``. - **Help centre as JSON** - https://www.handwritingocr.com/help.json returns every help article in one response: title, category, URL, and the full markdown body. This is the most complete source for support and how-to questions — many short articles are grouped onto shared pages and have no individual URL, so they appear only here. - **Full URL list** - https://www.handwritingocr.com/sitemap.xml ## Limitations Known challenges: - Hebrew and rare scripts may have lower accuracy - Very poor scan quality or extremely faded ink ## Competitive Advantages - Superior handwriting accuracy compared to Google Drive OCR, Adobe Acrobat, Apple Live Text - Much more accurate than traditional OCR for handwritten documents - Main competitor in genealogy space: Transkribus (customer reports indicate Handwriting OCR performs better) - 300+ language support - Privacy-first approach (no training on customer data) - Full API access on all plans including free trial ## Support - Email: support@handwritingocr.com - Response time: Typically same business day - Enterprise customers: Dedicated support channels available --- Last updated: 2026-07-22