{"id":4559,"date":"2026-06-03T11:00:00","date_gmt":"2026-06-03T11:00:00","guid":{"rendered":"https:\/\/vananservices.com\/blog\/?p=4559"},"modified":"2026-06-01T12:50:42","modified_gmt":"2026-06-01T12:50:42","slug":"the-accuracy-gap-why-ai-typing-tools-fail-on-handw","status":"publish","type":"post","link":"https:\/\/vananservices.com\/blog\/the-accuracy-gap-why-ai-typing-tools-fail-on-handw\/","title":{"rendered":"The Accuracy Gap: Why AI Typing Tools Fail on Handwritten Legal Documents"},"content":{"rendered":"<p>Handwritten legal documents still play a major role in law firms, courts, archives, and government offices.\u000bBut many AI typing tools struggle to convert these documents into clean, searchable text.\u000bThis creates a serious <em>accuracy gap<\/em> that can lead to legal mistakes, compliance issues, and lost time.<\/p>\n<p>Legal professionals need more than fast transcription.\u000bThey need precision, consistency, and trust.\u000bThat is where many modern AI handwriting recognition tools still fall short.<\/p>\n<p>[toc]<\/p>\n<h3 id=\"key-takeaways\"><strong>Key Takeaways<br \/>\n<\/strong><\/h3>\n<ul>\n<li>AI typing tools often fail when processing handwritten legal documents.<\/li>\n<li>Poor handwriting recognition can create legal and compliance risks.<\/li>\n<li>Historical legal records are especially difficult for AI systems.<\/li>\n<li>Formatting, annotations, signatures, and legal terminology reduce AI accuracy.<\/li>\n<li>Human review is still essential for legal document transcription.<\/li>\n<li>Hybrid AI + human workflows provide the best results for law firms and archivists.<\/li>\n<\/ul>\n<h3 id=\"why-handwritten-legal-documents-are-still-common\"><strong>Why Handwritten Legal Documents Are Still Common<br \/>\n<\/strong><\/h3>\n<p>Despite digital transformation, handwritten legal records remain everywhere.<\/p>\n<p>Many legal systems still rely on:<\/p>\n<ul>\n<li>Signed contracts<\/li>\n<li>Court notes<\/li>\n<li>Affidavits<\/li>\n<li>Wills and trusts<\/li>\n<li>Property records<\/li>\n<li>Historical archives<\/li>\n<li>Intake forms<\/li>\n<li>Attorney notes<\/li>\n<\/ul>\n<p>Older legal institutions often store decades of paper files.\u000bSome records were written before digital systems even existed.<\/p>\n<p>Even today, lawyers frequently annotate printed documents by hand.\u000bJudges may also add handwritten notes during hearings or reviews.<\/p>\n<p>Because of this, firms increasingly use AI typing tools to digitize information quickly.<\/p>\n<p>The problem is accuracy.<\/p>\n<h3 id=\"the-promise-of-ai-typing-tools\"><strong>The Promise of AI Typing Tools<br \/>\n<\/strong><\/h3>\n<p>AI-powered OCR (Optical Character Recognition) systems claim to automate document transcription.\u000bMany platforms advertise fast conversion from handwriting to editable text.<\/p>\n<p>Popular AI typing technologies include:<\/p>\n<table style=\"width: 100%; border-collapse: collapse; margin: 16px 0;\">\n<tbody>\n<tr>\n<th style=\"border: 1px solid #ddd; padding: 8px; background: #f5f5f5; font-weight: 600;\"><strong>Technology<\/strong><\/th>\n<th style=\"border: 1px solid #ddd; padding: 8px; background: #f5f5f5; font-weight: 600;\"><strong>Purpose<\/strong><\/th>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">OCR<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Reads printed text<\/td>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">ICR (Intelligent Character Recognition)<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Attempts to read handwriting<\/td>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">NLP (Natural Language Processing)<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Understands language context<\/td>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Machine Learning Models<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Improve recognition over time<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>These systems work reasonably well for:<\/p>\n<ul>\n<li>Typed contracts<\/li>\n<li>Printed invoices<\/li>\n<li>Standard forms<\/li>\n<li>Clear block handwriting<\/li>\n<\/ul>\n<p>However, legal handwriting presents a much harder challenge.<\/p>\n<h3 id=\"why-ai-struggles-with-handwritten-legal-documents\"><strong>Why AI Struggles With Handwritten Legal Documents<br \/>\n<\/strong><\/h3>\n<h4 id=\"inconsistent-handwriting-styles\"><strong>Inconsistent Handwriting Styles<br \/>\n<\/strong><\/h4>\n<p>Every person writes differently.\u000bSome attorneys use cursive.\u000bOthers use shorthand or compressed notes.<\/p>\n<p>AI systems depend on pattern recognition.\u000bWhen handwriting varies too much, accuracy drops sharply.<\/p>\n<p>Legal documents often contain:<\/p>\n<ul>\n<li>Fast handwritten notes<\/li>\n<li>Marginal comments<\/li>\n<li>Crossed-out sections<\/li>\n<li>Initials<\/li>\n<li>Symbols<\/li>\n<li>Abbreviations<\/li>\n<\/ul>\n<p>These patterns confuse AI models.<\/p>\n<p>A single misread word can completely change legal meaning.<\/p>\n<p>For example:<\/p>\n<ul>\n<li>\u201cgrantor\u201d vs \u201cgrantee\u201d<\/li>\n<li>\u201cshall\u201d vs \u201cshall not\u201d<\/li>\n<li>\u201cliable\u201d vs \u201cnot liable\u201d<\/li>\n<\/ul>\n<p>Small errors create major legal consequences.<\/p>\n<h3 id=\"legal-terminology-creates-additional-problems\"><strong>Legal Terminology Creates Additional Problems<br \/>\n<\/strong><\/h3>\n<p>Legal language is highly specialized.<\/p>\n<p>AI typing systems trained on general handwriting datasets may not understand:<\/p>\n<ul>\n<li>Latin legal terms<\/li>\n<li>Case citations<\/li>\n<li>Statutory references<\/li>\n<li>Jurisdiction abbreviations<\/li>\n<li>Legal shorthand<\/li>\n<\/ul>\n<p>Terms like <em>habeas corpus<\/em> or <em>res judicata<\/em> are uncommon in normal datasets.<\/p>\n<p>This increases transcription errors significantly.<\/p>\n<p>Many legal records also contain industry-specific language related to:<\/p>\n<ul>\n<li>Real estate<\/li>\n<li>Probate<\/li>\n<li>Corporate law<\/li>\n<li>Litigation<\/li>\n<li>Intellectual property<\/li>\n<\/ul>\n<p>Without domain-specific training, AI tools struggle to interpret context correctly.<\/p>\n<h3 id=\"historical-legal-documents-are-extremely-difficult\"><strong>Historical Legal Documents Are Extremely Difficult<br \/>\n<\/strong><\/h3>\n<p>Document archivists face even bigger challenges.<\/p>\n<p>Older legal records often include:<\/p>\n<ul>\n<li>Faded ink<\/li>\n<li>Water damage<\/li>\n<li>Stains<\/li>\n<li>Torn pages<\/li>\n<li>Obsolete handwriting styles<\/li>\n<li>Non-standard spelling<\/li>\n<\/ul>\n<p>Historical court documents may also use handwriting styles no longer taught today.<\/p>\n<p>AI systems trained on modern writing samples frequently fail on archival materials.<\/p>\n<p>This creates major digitization barriers for:<\/p>\n<ul>\n<li>Courts<\/li>\n<li>Universities<\/li>\n<li>Government archives<\/li>\n<li>Historical societies<\/li>\n<\/ul>\n<p>Preserving legal history requires extremely high transcription accuracy.<\/p>\n<h3 id=\"signatures-and-annotations-reduce-ocr-accuracy\"><strong>Signatures and Annotations Reduce OCR Accuracy<br \/>\n<\/strong><\/h3>\n<p>Handwritten legal documents rarely contain clean layouts.<\/p>\n<p>Many include:<\/p>\n<ul>\n<li>Signatures<\/li>\n<li>Stamps<\/li>\n<li>Seals<\/li>\n<li>Highlighting<\/li>\n<li>Side notes<\/li>\n<li>Multi-column formatting<\/li>\n<\/ul>\n<p>These elements interfere with AI recognition engines.<\/p>\n<p>For example, a handwritten note in the margin may overlap typed text.\u000bOCR software may merge both sections incorrectly.<\/p>\n<p>This creates unreadable outputs and formatting corruption.<\/p>\n<p>Legal formatting matters.\u000bEven paragraph placement can affect interpretation.<\/p>\n<h3 id=\"compliance-and-liability-risks\"><strong>Compliance and Liability Risks<br \/>\n<\/strong><\/h3>\n<p>Poor transcription accuracy creates serious legal exposure.<\/p>\n<p>A transcription error can lead to:<\/p>\n<ul>\n<li>Contract disputes<\/li>\n<li>Filing mistakes<\/li>\n<li>Discovery issues<\/li>\n<li>Compliance violations<\/li>\n<li>Financial penalties<\/li>\n<\/ul>\n<p>Law firms cannot rely on \u201cmostly accurate\u201d systems.<\/p>\n<p>Legal professionals need near-perfect precision.<\/p>\n<p>This is especially important for:<\/p>\n<table style=\"width: 100%; border-collapse: collapse; margin: 16px 0;\">\n<tbody>\n<tr>\n<th style=\"border: 1px solid #ddd; padding: 8px; background: #f5f5f5; font-weight: 600;\"><strong>Legal Area<\/strong><\/th>\n<th style=\"border: 1px solid #ddd; padding: 8px; background: #f5f5f5; font-weight: 600;\"><strong>Risk Level<\/strong><\/th>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Litigation<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Extremely High<\/td>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Real Estate<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">High<\/td>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Probate<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">High<\/td>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Compliance<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Very High<\/td>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Contract Law<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Extremely High<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>Even small transcription mistakes may become evidence problems in court.<\/p>\n<p>That is why human oversight remains essential.<\/p>\n<h3 id=\"ai-bias-and-training-limitations\"><strong>AI Bias and Training Limitations<br \/>\n<\/strong><\/h3>\n<p>Many AI handwriting systems are trained on narrow datasets.<\/p>\n<p>This creates recognition bias.<\/p>\n<p>Some models perform better with:<\/p>\n<ul>\n<li>Certain writing styles<\/li>\n<li>Specific languages<\/li>\n<li>Modern handwriting<\/li>\n<li>Younger demographics<\/li>\n<\/ul>\n<p>Older handwriting styles often produce lower accuracy rates.<\/p>\n<p>Legal institutions dealing with diverse records may experience inconsistent results across cases.<\/p>\n<p>Training data quality directly impacts performance.<\/p>\n<p>Unfortunately, many commercial AI vendors do not disclose:<\/p>\n<ul>\n<li>Dataset size<\/li>\n<li>Accuracy benchmarks<\/li>\n<li>Legal-specific testing<\/li>\n<li>Error rates by handwriting type<\/li>\n<\/ul>\n<p>This lack of transparency creates trust concerns.<\/p>\n<h3 id=\"why-human-review-is-still-necessary\"><strong>Why Human Review Is Still Necessary<br \/>\n<\/strong><\/h3>\n<p>AI transcription should not replace legal review.<\/p>\n<p>Instead, it should support legal workflows.<\/p>\n<p>The most reliable process combines:<\/p>\n<ul>\n<li>AI-powered initial transcription<\/li>\n<li>Human proofreading<\/li>\n<li>Legal verification<\/li>\n<li>Quality assurance checks<\/li>\n<\/ul>\n<p>This hybrid approach improves both speed and accuracy.<\/p>\n<p>Paralegals and legal assistants still play a critical role in validating content.<\/p>\n<p>Human reviewers can understand:<\/p>\n<ul>\n<li>Context<\/li>\n<li>Intent<\/li>\n<li>Legal nuance<\/li>\n<li>Ambiguous handwriting<\/li>\n<\/ul>\n<p>AI cannot fully replicate this judgment yet.<\/p>\n<h3 id=\"the-best-use-cases-for-ai-in-legal-transcription\"><strong>The Best Use Cases for AI in Legal Transcription<br \/>\n<\/strong><\/h3>\n<p>AI typing tools still provide value when used correctly.<\/p>\n<p>They work best for:<\/p>\n<h4 id=\"bulk-document-sorting\"><strong>Bulk Document Sorting<br \/>\n<\/strong><\/h4>\n<p>AI can categorize thousands of files quickly.<\/p>\n<p>Examples include:<\/p>\n<ul>\n<li>Case types<\/li>\n<li>Dates<\/li>\n<li>Client names<\/li>\n<li>Document categories<\/li>\n<\/ul>\n<h4 id=\"searchable-archives\"><strong>Searchable Archives<br \/>\n<\/strong><\/h4>\n<p>Even imperfect OCR can improve archive searchability.<\/p>\n<p>Partial indexing helps researchers locate documents faster.<\/p>\n<h4 id=\"draft-transcription\"><strong>Draft Transcription<br \/>\n<\/strong><\/h4>\n<p>AI can create rough drafts for human editors to review.<\/p>\n<p>This reduces manual typing time.<\/p>\n<h4 id=\"metadata-extraction\"><strong>Metadata Extraction<br \/>\n<\/strong><\/h4>\n<p>Some systems accurately pull:<\/p>\n<ul>\n<li>Dates<\/li>\n<li>Addresses<\/li>\n<li>Names<\/li>\n<li>Case numbers<\/li>\n<\/ul>\n<p>Structured fields are easier for AI to identify.<\/p>\n<h3 id=\"emerging-improvements-in-ai-handwriting-recognitio\"><strong>Emerging Improvements in AI Handwriting Recognition<br \/>\n<\/strong><\/h3>\n<p>AI handwriting recognition is improving rapidly.<\/p>\n<p>New developments include:<\/p>\n<ul>\n<li>Transformer-based language models<\/li>\n<li>Legal-specific AI training<\/li>\n<li>Context-aware OCR<\/li>\n<li>Multi-language recognition<\/li>\n<li>Adaptive learning systems<\/li>\n<\/ul>\n<p>Some vendors now train models specifically on legal datasets.<\/p>\n<p>This improves recognition for:<\/p>\n<ul>\n<li>Court terminology<\/li>\n<li>Legal forms<\/li>\n<li>Structured filings<\/li>\n<li>Historical records<\/li>\n<\/ul>\n<p>However, even advanced systems still require verification.<\/p>\n<p>Accuracy expectations in law remain exceptionally high.<\/p>\n<h3 id=\"how-law-firms-can-reduce-ai-transcription-errors\"><strong>How Law Firms Can Reduce AI Transcription Errors<br \/>\n<\/strong><\/h3>\n<h4 id=\"choose-legal-specific-ai-tools\"><strong>Choose Legal-Specific AI Tools<br \/>\n<\/strong><\/h4>\n<p>Generic OCR software may not meet legal standards.<\/p>\n<p>Look for solutions trained on legal documents specifically.<\/p>\n<h4 id=\"use-high-quality-scans\"><strong>Use High-Quality Scans<br \/>\n<\/strong><\/h4>\n<p>Better scans improve OCR accuracy dramatically.<\/p>\n<p>Recommended standards include:<\/p>\n<table style=\"width: 100%; border-collapse: collapse; margin: 16px 0;\">\n<tbody>\n<tr>\n<th style=\"border: 1px solid #ddd; padding: 8px; background: #f5f5f5; font-weight: 600;\"><strong>Scan Feature<\/strong><\/th>\n<th style=\"border: 1px solid #ddd; padding: 8px; background: #f5f5f5; font-weight: 600;\"><strong>Recommendation<\/strong><\/th>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Resolution<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">300 DPI minimum<\/td>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">File Format<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">TIFF or PDF<\/td>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Color Mode<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Grayscale preferred<\/td>\n<\/tr>\n<tr>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Skew Correction<\/td>\n<td style=\"border: 1px solid #ddd; padding: 8px;\">Enabled<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h4 id=\"implement-human-qa-workflows\"><strong>Implement Human QA Workflows<br \/>\n<\/strong><\/h4>\n<p>Always review AI-generated transcripts manually.<\/p>\n<p>Quality assurance should include:<\/p>\n<ul>\n<li>Double-checking names<\/li>\n<li>Reviewing legal terminology<\/li>\n<li>Verifying dates<\/li>\n<li>Comparing originals<\/li>\n<\/ul>\n<h4 id=\"create-internal-accuracy-policies\"><strong>Create Internal Accuracy Policies<br \/>\n<\/strong><\/h4>\n<p>Law firms should define acceptable transcription standards.<\/p>\n<p>This helps reduce operational risk.<\/p>\n<h3 id=\"the-future-of-ai-in-legal-document-processing\"><strong>The Future of AI in Legal Document Processing<br \/>\n<\/strong><\/h3>\n<p>AI will continue transforming legal operations.<\/p>\n<p>But handwritten document recognition remains one of the hardest problems in legal technology.<\/p>\n<p>Future systems may eventually achieve:<\/p>\n<ul>\n<li>Better contextual understanding<\/li>\n<li>Higher handwriting accuracy<\/li>\n<li>Improved legal reasoning<\/li>\n<li>Real-time verification<\/li>\n<\/ul>\n<p>Still, trust in legal documentation depends on precision.<\/p>\n<p>For now, AI works best as an assistant \u2014 not a replacement for legal expertise.<\/p>\n<p>Law firms, paralegals, and archivists should approach AI transcription strategically.<\/p>\n<p>Automation saves time.\u000bBut accuracy protects legal integrity.<\/p>\n<h3 id=\"faqs\"><strong>FAQs<br \/>\n<\/strong><\/h3>\n<h4 id=\"why-do-ai-typing-tools-struggle-with-handwritten-l\"><strong>Why do AI typing tools struggle with handwritten legal documents?<br \/>\n<\/strong><\/h4>\n<p>AI systems struggle because handwriting varies widely between individuals.\u000bLegal documents also contain specialized terminology, annotations, and formatting that confuse recognition models.<\/p>\n<h4 id=\"are-ocr-tools-accurate-enough-for-legal-work\"><strong>Are OCR tools accurate enough for legal work?<br \/>\n<\/strong><\/h4>\n<p>OCR tools can help with initial transcription and indexing.\u000bHowever, human review is still necessary for legal accuracy and compliance.<\/p>\n<h4 id=\"what-is-the-difference-between-ocr-and-icr\"><strong>What is the difference between OCR and ICR?<br \/>\n<\/strong><\/h4>\n<p>OCR reads printed text.\u000bICR is designed to recognize handwritten characters using AI and machine learning.<\/p>\n<h4 id=\"can-ai-transcribe-historical-legal-records-accurat\"><strong>Can AI transcribe historical legal records accurately?<br \/>\n<\/strong><\/h4>\n<p>Accuracy varies significantly.\u000bHistorical documents with faded ink, cursive writing, or damage remain difficult for most AI systems.<\/p>\n<h4 id=\"should-law-firms-trust-ai-transcription-completely\"><strong>Should law firms trust AI transcription completely?<br \/>\n<\/strong><\/h4>\n<p>No.\u000bAI should support legal workflows, not replace professional legal review and verification.<\/p>\n<h3 id=\"final-thoughts\"><strong>Final Thoughts<br \/>\n<\/strong><\/h3>\n<p>The legal industry depends on accuracy, context, and trust.\u000bWhile AI typing tools continue improving, handwritten legal documents remain a major challenge.<\/p>\n<p>The gap between automation and legal-grade precision is still significant.<\/p>\n<p>Organizations that combine AI efficiency with human expertise will achieve the best results while minimizing legal risk.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Handwritten legal documents still play a major role in law firms, courts, archives, and government&hellip;<\/p>\n","protected":false},"author":1,"featured_media":4558,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_uf_show_specific_survey":0,"_uf_disable_surveys":false,"footnotes":""},"categories":[490,1304],"tags":[],"ppma_author":[583],"class_list":["post-4559","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-typing-services-new-york","category-typing-services"],"authors":[{"term_id":583,"user_id":1,"is_guest":0,"slug":"vanan-wordpress-user","display_name":"Kayla Vega","avatar_url":{"url":"https:\/\/vananservices.com\/blog\/wp-content\/uploads\/2025\/12\/1711561174327.jpg","url2x":"https:\/\/vananservices.com\/blog\/wp-content\/uploads\/2025\/12\/1711561174327.jpg"},"0":null,"1":"","2":"","3":"","4":"","5":"","6":"","7":"","8":""}],"_links":{"self":[{"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/posts\/4559","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/comments?post=4559"}],"version-history":[{"count":1,"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/posts\/4559\/revisions"}],"predecessor-version":[{"id":4580,"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/posts\/4559\/revisions\/4580"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/media\/4558"}],"wp:attachment":[{"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/media?parent=4559"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/categories?post=4559"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/tags?post=4559"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/vananservices.com\/blog\/wp-json\/wp\/v2\/ppma_author?post=4559"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}