[{"data":1,"prerenderedAt":437},["ShallowReactive",2],{"blog":3},[4,204,347],{"id":5,"title":6,"author":7,"body":8,"date":194,"description":195,"extension":196,"meta":197,"navigation":198,"path":199,"seo":200,"sitemap":201,"stem":202,"__hash__":203},"blog\u002Fblog\u002Ffrom-wordpress-to-nuxt-ssg.md","From WordPress to Static: Rebuilding Our Website with Nuxt","Jani Golob",{"type":9,"value":10,"toc":187},"minimark",[11,15,18,23,26,55,63,67,70,137,142,149,154,160,163,177,181,184],[12,13,14],"p",{},"Our old website ran on WordPress. It worked, but it carried the usual baggage: a PHP runtime and database to keep patched, a theme layered with plugins, and render-blocking assets we didn't control. For a site that changes a few times a month, that's a lot of moving parts for no benefit.",[12,16,17],{},"So we rebuilt it.",[19,20,22],"h2",{"id":21},"the-new-stack","The New Stack",[12,24,25],{},"The site is now a statically generated Nuxt app:",[27,28,29,37,43,49],"ul",{},[30,31,32,36],"li",{},[33,34,35],"strong",{},"Nuxt with SSG:"," every page is pre-rendered to plain HTML at build time. No server-side rendering at request time, no origin compute at all.",[30,38,39,42],{},[33,40,41],{},"Nuxt Content:"," blog posts like this one are markdown files in the repo, parsed at build time.",[30,44,45,48],{},[33,46,47],{},"Tailwind CSS:"," styles are purged down to what the pages actually use.",[30,50,51,54],{},[33,52,53],{},"AWS CloudFront:"," the build output is uploaded to S3 and served from CloudFront edge locations. Every request is a cache hit on a static file.",[12,56,57,58,62],{},"Deployment is a git push. CodeBuild runs ",[59,60,61],"code",{},"nuxt generate"," and syncs the output — no servers, no database, nothing to patch.",[19,64,66],{"id":65},"the-numbers","The Numbers",[12,68,69],{},"Lighthouse (mobile), before and after:",[71,72,73,89],"table",{},[74,75,76],"thead",{},[77,78,79,83,86],"tr",{},[80,81,82],"th",{},"Metric",[80,84,85],{},"WordPress",[80,87,88],{},"Nuxt SSG",[90,91,92,104,115,126],"tbody",{},[77,93,94,98,101],{},[95,96,97],"td",{},"Performance score",[95,99,100],{},"49",[95,102,103],{},"91",[77,105,106,109,112],{},[95,107,108],{},"Largest Contentful Paint",[95,110,111],{},"7.6 s",[95,113,114],{},"3.1 s",[77,116,117,120,123],{},[95,118,119],{},"First Contentful Paint",[95,121,122],{},"2.7 s",[95,124,125],{},"2.4 s",[77,127,128,131,134],{},[95,129,130],{},"Cumulative Layout Shift",[95,132,133],{},"0.618",[95,135,136],{},"0.003",[12,138,139],{},[33,140,141],{},"Before — WordPress:",[12,143,144],{},[145,146],"img",{"alt":147,"src":148},"Lighthouse report for the WordPress site: performance score 49, LCP 7.6 s, CLS 0.618","\u002Fblog\u002Flighthouse-wordpress.webp",[12,150,151],{},[33,152,153],{},"After — Nuxt SSG on CloudFront:",[12,155,156],{},[145,157],{"alt":158,"src":159},"Lighthouse report for the Nuxt site: performance score 91, LCP 3.1 s, CLS 0.003","\u002Fblog\u002Flighthouse-nuxt.webp",[12,161,162],{},"The two metrics that moved the most are the ones users feel:",[27,164,165,171],{},[30,166,167,170],{},[33,168,169],{},"LCP dropped from 7.6 s to 3.1 s."," The hero content is in the initial HTML and the largest asset is an optimized WebP served from the edge, instead of being assembled by PHP and fetched from a distant origin.",[30,172,173,176],{},[33,174,175],{},"CLS dropped from 0.618 to 0.003."," The old theme loaded fonts and widgets that pushed content around as they arrived. Now layout is fixed at build time and nothing shifts after first paint.",[19,178,180],{"id":179},"why-it-matters","Why It Matters",[12,182,183],{},"Beyond the score: the attack surface is gone (static files have no admin panel to exploit), hosting costs are near zero, and the whole site is versioned in git — content included. Reviewing a blog post is a pull request, not a WYSIWYG editor.",[12,185,186],{},"If your site is mostly content that changes on human timescales, you probably don't need a CMS runtime. Build it once, put it on a CDN, and move on.",{"title":188,"searchDepth":189,"depth":189,"links":190},"",2,[191,192,193],{"id":21,"depth":189,"text":22},{"id":65,"depth":189,"text":66},{"id":179,"depth":189,"text":180},"2026-07-14","We replaced our WordPress site with a statically generated Nuxt app served from AWS CloudFront. The Lighthouse performance score went from 49 to 91.","md",{},true,"\u002Fblog\u002Ffrom-wordpress-to-nuxt-ssg",{"title":6,"description":195},{"loc":199},"blog\u002Ffrom-wordpress-to-nuxt-ssg","qBpY6VvqPrlmzXoWf2_vCPuyBSqESUCt77rpttQ-c2Y",{"id":205,"title":206,"author":207,"body":208,"date":339,"description":340,"extension":196,"meta":341,"navigation":198,"path":342,"seo":343,"sitemap":344,"stem":345,"__hash__":346},"blog\u002Fblog\u002Fbuilding-neurotko-multi-agent-platform.md","Building Neurotko: A Multi-Agent Platform for Real Business Processes","Simon Mavsar",{"type":9,"value":209,"toc":330},[210,213,216,220,223,226,258,261,265,268,271,274,278,281,284,287,291,294,297,301,304,307,311,314,318,321],[12,211,212],{},"Three things kept coming up in our client work: documents that needed reading, workflows that needed routing, and data that needed connecting across systems. We built one-off agents to handle each. At some point, we had to decide whether to keep building bespoke agents or build a platform.",[12,214,215],{},"We built the platform. We called it Neurotko.",[19,217,219],{"id":218},"what-neurotko-actually-is","What Neurotko Actually Is",[12,221,222],{},"Neurotko is a multi-agent system for business process automation. Not a chatbot builder, not a generic AI wrapper: a platform that orchestrates specialized agents, each with focused capabilities and tools, connected to the workflows and data sources your business already uses.",[12,224,225],{},"The agents we have built or are building:",[27,227,228,234,240,246,252],{},[30,229,230,233],{},[33,231,232],{},"Info Agent:"," answers questions from your content (documents, web pages, FAQs)",[30,235,236,239],{},[33,237,238],{},"Business Analyst:"," reads contracts, tenders, and offers; extracts key data; flags risk",[30,241,242,245],{},[33,243,244],{},"Offers Agent:"," reads incoming supplier offers and prepares outgoing proposals for clients",[30,247,248,251],{},[33,249,250],{},"Accounting Agent:"," reads invoices, extracts structured data, tracks payment obligations",[30,253,254,257],{},[33,255,256],{},"Monitor Agent:"," observability across all agents — usage, errors, performance in one place",[12,259,260],{},"Each agent has a narrow job. That is intentional.",[19,262,264],{"id":263},"the-design-decision-we-made-early","The Design Decision We Made Early",[12,266,267],{},"The tempting path with multi-agent systems is to build a general-purpose orchestrator that routes any task to any agent. We chose narrower scope on purpose.",[12,269,270],{},"Every agent in Neurotko has defined tools, a versioned system prompt, and a specific domain. The Monitor Agent does not write invoices. The Accounting Agent does not do contract analysis. Narrow agents are easier to test, easier to improve, and easier to trust.",[12,272,273],{},"The orchestration layer handles what agents cannot do alone: decomposing tasks that span multiple domains, routing exceptions to the right place, and keeping a coherent audit trail across channel boundaries (Slack, email, web chat).",[19,275,277],{"id":276},"integration-first-not-infrastructure-first","Integration-First, Not Infrastructure-First",[12,279,280],{},"Most enterprise AI projects start with infrastructure discussions. Neurotko starts with the workflow.",[12,282,283],{},"Agents plug into what you already use: Slack, email, PDF\u002FXLS\u002FDOCX documents, internal APIs. The goal is low-friction deployment: upload your content, connect the channel, go live. We have had tenants running in under 30 minutes.",[12,285,286],{},"This is possible because of how we handle tool use. Agents do not read entire files and output entire files. They use targeted tools: find-and-replace for edits, structured extraction for data, comparison tools for benchmarking offers. This keeps token usage predictable and avoids the truncation failures that plague naive document-handling agents.",[19,288,290],{"id":289},"prompt-engineering-as-a-discipline","Prompt Engineering as a Discipline",[12,292,293],{},"One thing that separates reliable agents from impressive demos is prompt discipline. Every agent in Neurotko has a versioned system prompt with explicit rules: what to do, what not to do, how to handle edge cases, how to format output.",[12,295,296],{},"We treat prompt changes like code changes: versioned, tested, and reviewed. We have banned entire categories of filler language that crept into early agent responses. We enforce silent execution with explicit change reports so humans can verify what the agent did without reading a wall of narration. These rules compound over time and the difference between v1 and v20 of a prompt is significant.",[19,298,300],{"id":299},"multi-tenancy-and-observability","Multi-Tenancy and Observability",[12,302,303],{},"Neurotko is designed for multi-tenancy from the start. Each tenant gets an isolated knowledge base, system prompt configuration, API keys, and audit logs. The dashboard handles tenant management, security controls, and usage insights.",[12,305,306],{},"Observability is not optional. We log what each agent processed, what decision it made, and why. This matters for two reasons: audits and continuous improvement. If an agent misclassifies a document, you need to know it happened and you need enough context to fix the underlying prompt or data.",[19,308,310],{"id":309},"cloud-or-on-premises","Cloud or On-Premises",[12,312,313],{},"For clients in regulated industries, where data is processed is non-negotiable. Neurotko runs on both: cloud-hosted for fast deployment, or fully on-premises where the LLM stack never leaves your infrastructure. The agent architecture does not change. The deployment target does.",[19,315,317],{"id":316},"where-we-are","Where We Are",[12,319,320],{},"Neurotko is live internally and with beta clients. The Info Agent is the first module available: an AI assistant that answers questions from your content, deployable as a website chatbot, Slack bot, or Telegram assistant.",[12,322,323,324,329],{},"We are in internal beta. More modules are in development. If you want to be among the first to try it, or want to discuss deploying Neurotko for your business, reach out at ",[325,326,328],"a",{"href":327},"mailto:info@neurarod.com","info@neurarod.com",".",{"title":188,"searchDepth":189,"depth":189,"links":331},[332,333,334,335,336,337,338],{"id":218,"depth":189,"text":219},{"id":263,"depth":189,"text":264},{"id":276,"depth":189,"text":277},{"id":289,"depth":189,"text":290},{"id":299,"depth":189,"text":300},{"id":309,"depth":189,"text":310},{"id":316,"depth":189,"text":317},"2026-07-08","Most AI tooling is designed to impress in demos, not to run reliably in production. We built Neurotko to close that gap — a multi-agent platform focused on real workflows, real integrations, and measurable results.",{},"\u002Fblog\u002Fbuilding-neurotko-multi-agent-platform",{"title":206,"description":340},{"loc":342},"blog\u002Fbuilding-neurotko-multi-agent-platform","TlsBpv6QRfwv7ML8MU27mRbhkrl0RT1o7VduCetadsw",{"id":348,"title":349,"author":350,"body":351,"date":429,"description":430,"extension":196,"meta":431,"navigation":198,"path":432,"seo":433,"sitemap":434,"stem":435,"__hash__":436},"blog\u002Fblog\u002Fthe-future-of-cybersecurity-with-artificial-intelligence.md","The Future of Cybersecurity with Artificial Intelligence","Gregor Doltar",{"type":9,"value":352,"toc":419},[353,356,359,363,366,369,373,378,381,384,388,391,394,398,401,404,407,411,416],[12,354,355],{},"The NeuraRod team attended IBM Security Days 2026, where we had the opportunity to explore the latest trends, technologies, and solutions in cybersecurity and artificial intelligence. The event offered valuable insight into the direction modern security systems are taking and the growing role of AI in protecting data, infrastructure, and business processes.",[12,357,358],{},"We are especially pleased to see that the importance of cybersecurity and digital sovereignty is gaining stronger recognition at the national level. Topics such as system resilience, critical infrastructure protection, and the secure use of artificial intelligence are also being actively followed by URSIV, the Government Information Security Office of the Republic of Slovenia.",[19,360,362],{"id":361},"a-look-into-the-future-of-security","A Look Into the Future of Security",[12,364,365],{},"The event showcased a wide range of IBM security solutions, along with innovations from partner companies that complement the modern cybersecurity ecosystem. One of the key highlights was IBM QRadar, a leading SIEM solution that enables organisations to centrally monitor security events, detect threats faster, and respond to incidents more effectively.",[12,367,368],{},"The message of the event was clear: the future of cybersecurity will be built on intelligent automation and the effective use of artificial intelligence.",[19,370,372],{"id":371},"solutions-that-caught-our-attention","Solutions That Caught Our Attention",[374,375,377],"h3",{"id":376},"kyborg-by-kybm-ai-powered-incident-response","Kyborg by KYBM: AI-Powered Incident Response",[12,379,380],{},"One of the most interesting presentations was Kyborg, a platform that integrates with the IBM QRadar SIEM system. What stood out to us in particular was the ability to use both local, on-premises AI models and cloud-based AI models. This gives organisations the flexibility to keep sensitive security data within their own infrastructure.",[12,382,383],{},"This direction strongly aligns with how we see the future of NeuraRod: artificial intelligence that operates securely, reliably, and under the full control of the organisation.",[374,385,387],{"id":386},"ibm-vault-centralized-management-of-security-secrets","IBM Vault: Centralized Management of Security Secrets",[12,389,390],{},"Another important solution presented at the event was IBM Vault, which focuses on the centralised management of sensitive information such as API keys, passwords, and certificates.",[12,392,393],{},"As AI agents and automation become an increasingly important part of business processes, secure access and secrets management are becoming essential. In the development of NeuraRod agents, we also see alignment with Vault-based approaches as an important foundation for enterprise-grade security.",[19,395,397],{"id":396},"what-this-means-for-slovenian-companies","What This Means for Slovenian Companies",[12,399,400],{},"IBM Security Days 2026 clearly showed that organisations are looking for ways to safely integrate artificial intelligence into existing IT and security systems without losing control over their data.",[12,402,403],{},"This is relevant not only for large enterprises, but also for public institutions, where security, compliance, and data governance are becoming increasingly important. Slovenia is actively strengthening its national cybersecurity resilience framework and deepening cooperation at the European level.",[12,405,406],{},"At NeuraRod, we believe the future lies in AI agents that work directly with company data, within the organisation’s own infrastructure, and under the full control of the user.",[19,408,410],{"id":409},"thank-you-to-the-organisers-and-partners","Thank You to the Organisers and Partners",[12,412,413],{},[33,414,415],{},"We would like to thank the KYBM team and the organisers of IBM Security Days for the invitation and the opportunity to take part in such an important professional event.",[12,417,418],{},"We are proud to have been part of a community that is actively shaping the future of cybersecurity and artificial intelligence.",{"title":188,"searchDepth":189,"depth":189,"links":420},[421,422,427,428],{"id":361,"depth":189,"text":362},{"id":371,"depth":189,"text":372,"children":423},[424,426],{"id":376,"depth":425,"text":377},3,{"id":386,"depth":425,"text":387},{"id":396,"depth":189,"text":397},{"id":409,"depth":189,"text":410},"2026-06-02","Insights from IBM Security Days 2026 on how intelligent automation and AI are shaping the future of cybersecurity, from QRadar SIEM to secure on-premises AI models and secrets management.",{},"\u002Fblog\u002Fthe-future-of-cybersecurity-with-artificial-intelligence",{"title":349,"description":430},{"loc":432},"blog\u002Fthe-future-of-cybersecurity-with-artificial-intelligence","ooitY39zTh0AF6GOjt1WcghBqjXSTqBXQhbGYxY4INM",1784104548272]