{"id":8886,"date":"2026-04-08T12:47:59","date_gmt":"2026-04-08T12:47:59","guid":{"rendered":"https:\/\/www.neelnetworks.com\/blog\/?p=8886"},"modified":"2026-04-08T14:08:08","modified_gmt":"2026-04-08T14:08:08","slug":"agentic-ai-web-design-2026","status":"publish","type":"post","link":"https:\/\/www.neelnetworks.com\/blog\/agentic-ai-web-design-2026\/","title":{"rendered":"Agentic AI in Web Design: How Websites Are Starting to Manage Themselves"},"content":{"rendered":"
\n

\"Agentic<\/p>\n

For the past decade, AI’s role in web design has been reactive. You asked, it responded. You typed a prompt, it generated an image. You uploaded a brief, it produced copy. The AI waited for instructions. It had no initiative, no awareness of what the website needed, no ability to observe and act without being told to.<\/p>\n

That model is ending.<\/p>\n

In 2026, a new category of AI \u2014 agentic AI \u2014 is beginning to reshape what a website actually does. Instead of waiting for a human to observe a problem and instruct an AI tool to address it, agentic AI observes the website itself, identifies opportunities, makes decisions, and takes actions. Layouts shift based on what the data says works. Content adapts based on who is reading it. A\/B tests run and resolve without a human reviewing the results. Broken links get flagged, tagged, and routed for repair. The website, in a meaningful sense, begins to participate in its own management.<\/p>\n

This is not a distant prediction. It is happening now \u2014 in the tools web designers use, in the platforms enterprise companies are building, and in the AI-native features rolling out on major website platforms. This guide explains what agentic AI actually is, how it differs from the AI tools you are already familiar with, what it is currently doing in web design, and what it means practically for businesses planning their digital presence in 2026.<\/p>\n

What Is Agentic AI \u2014 And How Does It Differ From the AI You Already Know?<\/h2>\n

The AI tools most people are familiar with in 2026 are what researchers call “responsive” or “reactive” AI. ChatGPT, Gemini, Claude, Midjourney, Copilot \u2014 you give them an input, they produce an output. They are extraordinarily capable at what they do, but they are fundamentally passive. They do not initiate. They do not observe the world and decide that something needs to be done. They wait.<\/p>\n

Agentic AI is different in a foundational way. An AI agent has a goal, the ability to plan a sequence of actions to achieve that goal, access to tools that allow it to take those actions in the real world, and the ability to observe the results of those actions and adapt its next steps accordingly. It operates in a loop \u2014 observe, decide, act, observe, decide, act \u2014 until the goal is achieved or until it determines it cannot proceed without human input.<\/p>\n

\n

A simple illustration of the difference:<\/strong><\/p>\n

Reactive AI:<\/strong> You notice your website’s contact page has a high bounce rate. You ask an AI tool: “Write me five improved versions of this contact page headline.” It produces five headlines. You choose one and manually implement it.<\/p>\n

Agentic AI:<\/strong> The agent observes that the contact page has a high bounce rate. It identifies that visitors are scrolling past the form without engaging. It generates and deploys three headline variants as an A\/B test. It monitors results over 72 hours. It identifies the winning variant and implements it. It logs the change and the data behind it. It notifies you with a summary. You were never asked to do anything.<\/p>\n<\/div>\n

The distinction matters enormously for web design because websites are complex, dynamic systems with thousands of variables that benefit from continuous monitoring and optimisation \u2014 far more than any human team can manage manually at scale.<\/p>\n

\"Reactive<\/p>\n

The fundamental shift: reactive AI waits for human instruction; agentic AI observes, plans, acts, and adapts in a continuous autonomous cycle \u2014 with human oversight at the boundaries rather than at every step.<\/p>\n

Why 2026 Is the Year Agentic AI Enters Web Design at Scale<\/h2>\n

The concept of AI agents is not new \u2014 it has existed in academic AI research for decades. What is new in 2026 is the convergence of three conditions that make agentic AI practically deployable in web design contexts:<\/p>\n

Condition 1: Large Language Models That Can Reason and Plan<\/h3>\n

The current generation of large language models \u2014 GPT-4o, Claude 3.5, Gemini 1.5 Pro and their successors \u2014 are capable of multi-step reasoning at a level that previous models were not. They can receive a goal, decompose it into a sequence of sub-tasks, evaluate which tools to use for each sub-task, and adapt their approach based on intermediate results. This reasoning capability is the cognitive engine that makes useful agents possible \u2014 without it, agents either fail on complex tasks or produce outputs that require heavy human correction.<\/p>\n

Condition 2: Tool Access and Real-World Action<\/h3>\n

Recent infrastructure developments allow LLMs to use tools \u2014 to browse the web, execute code, read and write files, call APIs, and interact with software interfaces. The Model Context Protocol (MCP), released by Anthropic in late 2024 and rapidly adopted across the AI ecosystem, has standardised how AI models connect to external tools and data sources. This standardisation has dramatically accelerated the development of agents that can take meaningful actions in web design workflows rather than simply generating text about what those actions should be.<\/p>\n

Condition 3: Platform Integration and Production Deployment<\/h3>\n

Major web design platforms \u2014 Webflow, Framer, Wix, and even WordPress through third-party plugins \u2014 have begun integrating agentic capabilities into their core product rather than leaving them as external add-ons. Figma’s AI features now include agentic functions that can generate and iterate on designs based on specified goals. These integrations bring agentic AI into the design workflow at the point where web designers are already working, removing the adoption barrier of learning an entirely new toolset.<\/p>\n

What Agentic AI Is Actually Doing on Websites Right Now<\/h2>\n

Rather than describing theoretical future capabilities, here is a grounded account of what agentic AI is concretely doing on business websites in 2026.<\/p>\n

Autonomous Layout Optimisation<\/h3>\n

Platforms like Webflow and newer AI-native website builders are beginning to offer agents that monitor user behaviour data \u2014 heatmaps, scroll depth, click patterns, session recordings \u2014 and automatically propose or implement layout changes based on that data. An agent might identify that 60% of mobile visitors on a service page scroll past the CTA button without seeing it (because the button is positioned below an image that compresses on mobile). The agent repositions the button, monitors the change’s impact on the conversion rate, and either keeps it or reverts based on the results.<\/p>\n

This closes the loop between analytics insight and implementation action \u2014 a loop that has historically required a human to sit in the middle, reviewing data, briefing a designer, and waiting for the change to be deployed.<\/p>\n

Real-Time Content Personalisation<\/h3>\n

AI-powered personalisation has existed for years in large eCommerce platforms like Amazon and Netflix. In 2026, agentic personalisation is becoming accessible to mid-market and smaller businesses through platforms like Uniform, Ninetailed, and newer AI website tools. An agent monitors each visitor’s behaviour \u2014 which pages they visit, which content they engage with, where they came from, whether they are a returning visitor \u2014 and dynamically serves personalised versions of key pages. A visitor arriving from a LinkedIn ad for a specific service sees that service featured prominently. A returning visitor who previously browsed pricing sees a targeted offer. This happens in real time, at scale, without a human constructing each personalised variant manually.<\/p>\n

Automated A\/B Testing and Conversion Optimisation<\/h3>\n

Traditional A\/B testing requires a human to: identify a hypothesis, design the variant, configure the test, wait for statistical significance (often weeks), analyse the results, and implement the winner. Agentic A\/B testing compresses this loop. An agent monitors conversion metrics, identifies pages with below-benchmark performance, generates multiple variants based on CRO best practices, deploys them as an automated test, monitors results, determines statistical significance, implements the winner, and documents the learning \u2014 all without requiring a human to manage each step. Google’s own automated A\/B testing tools, Optimizely’s AI features, and VWO’s Copilot are all moving in this direction.<\/p>\n

Intelligent SEO Monitoring and Action<\/h3>\n

SEO agents are monitoring Search Console data, identifying pages where rankings are declining, diagnosing likely causes (thin content, missing schema, slow load time, declining click-through rate), and either implementing fixes automatically (updating meta data, adding schema markup, flagging technical issues for developer attention) or presenting prioritised action recommendations with the supporting data. Tools like Semrush’s AI features and newer SEO agent platforms are making this practically deployable for businesses beyond enterprise scale.<\/p>\n

Accessibility Auditing and Remediation<\/h3>\n

Automated accessibility scanning is not new. What is new is agents that not only identify WCAG violations but propose and in some cases implement fixes \u2014 generating descriptive alt text for images lacking it, adjusting colour contrast values that fail ratio requirements, restructuring heading hierarchies that are illogical for screen readers. UserWay, accessiBe, and similar platforms have been evolving in this direction for several years, and the agentic layer is now significantly more capable and accurate than earlier automated tools.<\/p>\n

Chatbots That Actually Resolve Problems<\/h3>\n

The website chatbot of 2020 was a rigid decision tree that frustrated more visitors than it helped. The agentic chatbot of 2026 is a different tool entirely. Connected to your product database, your booking system, your support ticket platform, and your knowledge base, it can: answer highly specific product questions by querying live inventory data, book appointments directly in your calendar system, create and update support tickets, process simple returns, and know when a query genuinely needs human intervention and route it efficiently. The distinction from a simple chatbot is that the agent takes actions \u2014 it does not just provide information.<\/p>\n

Agentic AI in Design Tools \u2014 Where Designers Are Already Experiencing It<\/h2>\n

\"Agentic<\/p>\n

In 2026’s leading design tools, AI agents are embedded directly into the design workflow \u2014 suggesting improvements, generating variants, and catching accessibility issues in real time rather than after the fact.<\/p>\n

Figma’s AI Features<\/h3>\n

Figma’s 2025 and 2026 AI updates have moved significantly toward agentic functionality. Figma Make \u2014 their AI-powered design assistant \u2014 can now receive a goal-level instruction (“redesign this pricing section to improve clarity and conversion”) and generate multiple complete layout variants, apply brand styles automatically, and iterate based on feedback. The AI is not just generating elements \u2014 it is understanding the goal, planning an approach, generating solutions, and refining based on stated criteria. Figma’s own 2025 AI report found that 51% of Figma users working on AI products are now building agents, compared to 21% the previous year.<\/p>\n

Webflow’s AI Capabilities<\/h3>\n

Webflow has integrated AI generation into its visual editor \u2014 allowing designers to describe sections, components, and interactions in natural language and have them built in the canvas. More significantly, Webflow’s AI can audit existing designs for accessibility compliance, suggest structural improvements for SEO, and generate responsive variants. The direction is clearly toward an AI assistant that co-designs rather than simply generates on demand.<\/p>\n

Framer’s AI Design System<\/h3>\n

Framer has been aggressive in integrating AI generation at every level of the design process \u2014 AI-generated layouts, AI-suggested interactions, and AI-powered content population from data sources. Framer’s AI can take a text description of a website and produce a complete, interactive prototype with realistic content, which designers then refine rather than build from scratch.<\/p>\n

What Agentic AI Means for Web Designers and Developers<\/h2>\n

The question every designer and developer asks when agentic AI comes up is the obvious one: does this replace us? The honest answer in 2026 is: some specific tasks, yes. Many more, no. And some entirely new categories of work, it creates.<\/p>\n

Tasks That Agents Are Beginning to Replace<\/h3>\n
    \n
  • Routine A\/B test setup, monitoring, and result analysis<\/li>\n
  • Basic SEO meta data updates and schema markup additions<\/li>\n
  • Accessibility scanning and straightforward remediation (alt text generation, contrast fixes)<\/li>\n
  • Generating initial layout variants for design review<\/li>\n
  • Populating designs with realistic content from a content source<\/li>\n
  • Performance monitoring and basic optimisation reporting<\/li>\n<\/ul>\n

    Tasks That Agents Cannot Yet Replicate<\/h3>\n
      \n
    • Understanding a client’s brand positioning, audience nuances, and competitive context at the depth a skilled designer brings<\/li>\n
    • Creative direction \u2014 the vision for what a website should feel like and why<\/li>\n
    • Complex custom development requiring business logic understanding and system integration expertise<\/li>\n
    • Client relationship management, project strategy, and the professional judgement that comes from years of experience<\/li>\n
    • Designing for genuinely novel problems that have no precedent in training data<\/li>\n
    • Quality assurance that requires understanding context and intent rather than pattern-matching against rules<\/li>\n<\/ul>\n

      New Work That Agents Are Creating<\/h3>\n
        \n
      • Agent configuration and orchestration<\/strong> \u2014 setting up, connecting, and managing AI agents is itself a growing skill area<\/li>\n
      • AI output curation and quality control<\/strong> \u2014 reviewing and selecting from agent-generated variants is a design skill<\/li>\n
      • Data interpretation and strategy<\/strong> \u2014 deciding which insights from agent monitoring should drive design decisions<\/li>\n
      • Prompt engineering for design<\/strong> \u2014 writing effective goal-level instructions that produce useful agent outputs<\/li>\n<\/ul>\n

        The practical shift for web designers and developers in 2026 is not elimination \u2014 it is elevation. Agents handle more of the repetitive, rule-based, data-processing work. Designers spend more time on strategy, creative direction, complex problem-solving, and client relationships. The floor of what a small design team can produce and maintain rises significantly.<\/p>\n

        What Agentic AI Means for Businesses Commissioning Websites<\/h2>\n

        If you are a business owner rather than a designer, agentic AI changes the website conversation in several practical ways:<\/p>\n

        Websites That Improve After Launch<\/h3>\n

        Traditionally, a website was at its best on launch day and slowly drifted from optimal as the business evolved, content aged, and user behaviour changed but the site did not. Agentic AI makes it realistic for a website to improve continuously after launch \u2014 optimising conversion elements, personalising content, monitoring SEO health, and catching accessibility issues \u2014 without requiring a new design project every time an improvement opportunity is identified.<\/p>\n

        Raising the Bar for What You Should Expect<\/h3>\n

        When evaluating web design agencies and partners in 2026, agentic AI capability is a legitimate differentiator to assess. Agencies that are integrating agentic tools into their design and optimisation workflows can deliver ongoing value that agencies operating purely with traditional tools cannot match at the same cost. Ask prospective partners how they use AI in their ongoing website management and optimisation work \u2014 the quality and specificity of the answer tells you a great deal about how current their practice is.<\/p>\n

        The Importance of Data Foundations<\/h3>\n

        Agentic AI on your website is only as good as the data it has access to. An agent that cannot read your Google Analytics 4, your Search Console, your heatmap data, and your conversion events cannot make meaningful optimisation decisions. Before deploying any agentic layer on your website, ensure your data foundations are correct: GA4 properly configured with conversion tracking, Google Search Console verified and monitored, and session recording tools active on key conversion pages.<\/p>\n

        The Risks and Limitations That Matter<\/h2>\n
        \n

        Honest assessment:<\/strong> Agentic AI in web design is genuinely transformative \u2014 and it is also genuinely early. The capabilities described in this guide exist and are being deployed. They are also imperfect, occasionally unpredictable, and in some cases actively harmful when deployed without appropriate human oversight. Here is what to watch for.<\/p>\n<\/div>\n

        Brand Drift Without Human Oversight<\/h3>\n

        An agent optimising for conversion rate will systematically move toward whatever layouts and copy perform best in A\/B tests \u2014 regardless of whether those layouts and that copy align with brand positioning, brand voice, or long-term brand building objectives. Conversion rate optimisation and brand integrity can be in tension. An agent without brand constraints and human review can optimise your website into something that converts slightly better in the short term but erodes the brand differentiation that drives long-term business value.<\/p>\n

        Optimising for the Wrong Metric<\/h3>\n

        An agent optimises for the metric it is given. If it is given “reduce bounce rate” as its goal, it will find ways to reduce bounce rate \u2014 which may or may not correlate with the business outcomes you actually care about (enquiries, sales, qualified leads). The most dangerous agentic deployments are those where the goal metric is imprecisely defined or where the agent’s optimisation of a proxy metric diverges from the underlying business objective.<\/p>\n

        Hallucination and Factual Errors at Scale<\/h3>\n

        AI agents generating website content \u2014 product descriptions, FAQ answers, service descriptions \u2014 can hallucinate: confidently producing text that is factually incorrect. At scale, without human review of every output, incorrect information can propagate across a website faster than it can be caught. Any agentic content generation workflow requires human review protocols, particularly for factual claims, pricing information, and any content in regulated categories.<\/p>\n

        Privacy and Data Compliance<\/h3>\n

        AI personalisation agents typically collect and process significant visitor behavioural data. In markets with strong privacy regulation \u2014 GDPR in Europe, UK GDPR post-Brexit, and similar frameworks in other markets \u2014 the data processing that personalisation agents perform must be compliant with consent requirements, data minimisation principles, and data residency rules. Deploying AI personalisation without proper GDPR compliance review is a legal risk.<\/p>\n

        Frequently Asked Questions About Agentic AI in Web Design<\/h2>\n\n\n\n\n\n\n\n\n\n
        What is agentic AI and how is it different from regular AI tools?<\/td>\nAgentic AI refers to AI systems that can pursue goals autonomously \u2014 observing their environment, planning sequences of actions, using tools to execute those actions, and adapting based on results \u2014 without requiring a human instruction at every step. Regular AI tools (ChatGPT, Gemini, Midjourney) are reactive: they wait for a human prompt, produce an output, and stop. An agentic AI observes that a website’s contact page has a declining conversion rate, identifies likely causes from behavioural data, generates and deploys A\/B test variants, monitors results, implements the winner, and logs the outcome \u2014 all without a human asking it to do each step. The difference is initiative: agentic AI acts, not just responds.<\/td>\n<\/tr>\n
        What can agentic AI currently do on a business website?<\/td>\nIn 2026, agentic AI can practically do the following on business websites: autonomously run A\/B tests on conversion-critical pages and implement winning variants; personalise website content in real time based on visitor behaviour and source; monitor SEO metrics and flag or implement improvements to meta data, schema markup, and content; audit websites for accessibility violations and generate remediation for straightforward issues (alt text, contrast, heading structure); power chatbots that take real actions (booking appointments, creating support tickets, querying live inventory) rather than just providing information; and monitor Core Web Vitals and page performance with proactive alerting on degradation. The most advanced deployments combine several of these capabilities through orchestrated agent workflows.<\/td>\n<\/tr>\n
        Is agentic AI going to replace web designers?<\/td>\nNo \u2014 but it is changing which parts of web design work require human skill. Agentic AI is beginning to handle routine, rule-based, data-processing tasks: A\/B test management, basic SEO updates, accessibility scanning, initial layout variant generation, and performance monitoring. These tasks will require less human time. The work that agents cannot replicate \u2014 creative direction, brand strategy, complex custom development, client relationship management, and design judgment in genuinely novel contexts \u2014 remains human work. The practical effect is that skilled designers and developers can produce more with less repetitive task overhead. The floor of what a small design team can accomplish rises. The value of strategic and creative expertise rises because agents increasingly handle the execution of well-specified tasks, making the specification of those tasks the higher-value skill.<\/td>\n<\/tr>\n
        What is the difference between an AI agent and an AI chatbot?<\/td>\nA traditional chatbot follows a decision tree or generates text responses \u2014 it provides information and answers questions, but it does not take actions in the world. An AI agent built on a large language model can use tools to take real actions: query a live database, book an appointment in a calendar system, create a support ticket, update a record, or trigger a workflow in a connected platform. The practical difference for website visitors is significant: a chatbot tells you what your order status is; an AI agent looks it up in real time, and if there is a problem, raises it with the relevant team and updates you automatically. The agentic chatbot resolves the issue; the traditional chatbot describes it.<\/td>\n<\/tr>\n
        Which web design tools have agentic AI features in 2026?<\/td>\nThe leading design and website tools with meaningful agentic AI capabilities in 2026 include: Figma (Figma Make for goal-level design generation and iteration), Webflow (AI design assistant for layout generation and accessibility auditing), Framer (AI-powered layout and content generation from natural language descriptions), Wix (AI site builder with ongoing AI suggestions), and Shopify (Shopify Sidekick for store management and Shopify Magic for content generation). In the optimisation space, Google’s AI-driven A\/B testing, Optimizely’s Copilot, and VWO’s AI features are bringing agentic optimisation to mid-market businesses. SEO tools including Semrush and Ahrefs are integrating agentic monitoring and recommendation features into their core platforms.<\/td>\n<\/tr>\n
        What are the risks of using agentic AI on a business website?<\/td>\nThe main risks of agentic AI on business websites are: brand drift, where an agent optimising for a metric like conversion rate moves website content and design away from brand positioning without human awareness; optimising for the wrong metric, where a precisely specified but imperfectly chosen goal leads the agent to improve a proxy measure without improving actual business outcomes; factual hallucination in AI-generated content, where agents produce incorrect information at scale faster than human review can catch it; data privacy compliance issues, where AI personalisation tools process visitor data in ways that may not comply with GDPR or other applicable regulations; and loss of human oversight, where agents make enough autonomous changes that the website evolves away from what the business intended. Mitigating these risks requires: clear goal definition, human review protocols, brand guardrails built into agent instructions, and regular oversight of what agents are actually doing.<\/td>\n<\/tr>\n
        How should a small business approach agentic AI for their website in 2026?<\/td>\nFor small businesses, the most practical approach to agentic AI in 2026 is to start with the specific, well-bounded agent capabilities available through the platforms you already use rather than deploying complex multi-agent systems. Concretely: if you use Shopify, activate Shopify Sidekick for store management assistance and Shopify Magic for product description generation. If you use a website builder like Webflow or Wix, enable their AI design suggestions and accessibility auditing. Install an AI-powered chatbot (Tidio, Intercom) for after-hours lead capture. Ensure your analytics are properly configured so any future agent-driven optimisation has good data to work with. Graduate to more sophisticated agentic optimisation (AI-driven A\/B testing, personalisation) once you have confirmed that basic data infrastructure is sound and you have a clear understanding of the goals you want agents to pursue.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

        \"Agentic<\/p>\n

        The practical output of agentic AI on a well-configured business website: measurable improvements in conversion, SEO, and visitor experience \u2014 delivered continuously, with human oversight at the strategic level rather than the execution level.<\/p>\n

        \n

        Want to Know How Agentic AI can Improve Your Business Website in 2026?<\/strong><\/p>\n

        Neel Networks is actively integrating agentic AI tools into our web design, development, and ongoing optimization services. Talk to our team about how AI-driven optimization can improve your website\u2019s performance, SEO, and conversion rates\u2014with proper human oversight built in at every step.<\/p>\n

        Web Design Services<\/a> Free AI Consultation<\/a> WhatsApp Us<\/a>\n<\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":"

        For the past decade, AI’s role in web design has been reactive. You asked, it responded. You typed a prompt, it generated an image. You uploaded a brief, it produced copy. The AI waited for instructions. It had no initiative, no awareness of what the website needed, no ability to observe and act without being […]<\/p>\n","protected":false},"author":1,"featured_media":8889,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[453,1],"tags":[],"class_list":["post-8886","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-future-tech","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/www.neelnetworks.com\/blog\/wp-json\/wp\/v2\/posts\/8886","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.neelnetworks.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.neelnetworks.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.neelnetworks.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.neelnetworks.com\/blog\/wp-json\/wp\/v2\/comments?post=8886"}],"version-history":[{"count":12,"href":"https:\/\/www.neelnetworks.com\/blog\/wp-json\/wp\/v2\/posts\/8886\/revisions"}],"predecessor-version":[{"id":8905,"href":"https:\/\/www.neelnetworks.com\/blog\/wp-json\/wp\/v2\/posts\/8886\/revisions\/8905"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.neelnetworks.com\/blog\/wp-json\/wp\/v2\/media\/8889"}],"wp:attachment":[{"href":"https:\/\/www.neelnetworks.com\/blog\/wp-json\/wp\/v2\/media?parent=8886"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.neelnetworks.com\/blog\/wp-json\/wp\/v2\/categories?post=8886"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.neelnetworks.com\/blog\/wp-json\/wp\/v2\/tags?post=8886"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}