Technical SEO is the part of search optimisation which developers control: whether search engines can crawl, render, index and understand every page. A 2026 checklist follows the order of Googlebot. Step 1 is crawling, with correct status codes, robots.txt and an XML sitemap. Step 2 is indexing, with canonical and hreflang tags for duplicate and multilingual URLs. Step 3 is rendering, with complete HTML from the server. Step 4 is speed, measured with Core Web Vitals at the 75th percentile of real visits. According to the HTTP Archive Web Almanac 2025, only 48% of mobile sites pass all three Core Web Vitals. Finally, structured data makes every page easier to understand. The same foundation decides eligibility for AI features in Google Search, according to Google. For example, a business website with 20 pages built with Next.js can pass every check.
What is technical SEO for developers?
Technical SEO is the work which makes a website easy for search engines to crawl, render, index and understand. Content and links attract visitors, but technical SEO decides whether Google can process the pages at all. Googlebot handles JavaScript web apps in three phases, according to Google: crawling, rendering and indexing. A developer influences every phase. Step 1 is the server, which must return correct status codes and fast responses. Step 2 is the HTML, which must contain real links, titles and content without waiting for client-side scripts. Finally, signals such as canonical tags, hreflang and structured data must tell one consistent story. In practice, most technical SEO problems come from framework defaults, for example a missing canonical tag or a staging site which accidentally got indexed. The checklist below follows the same order as Googlebot in 2026, so a developer can audit a site built with Next.js, WordPress or any other stack.
How do you make sure Google can crawl and index every page?
Crawlability is the ability of a crawler to discover and fetch every important URL on a site. Step 1 is internal links. Googlebot finds new URLs by parsing the href attribute of HTML links, according to Google. Navigation built only from buttons with click handlers can therefore hide pages. Step 2 is robots.txt, which Google describes as a tool for crawl management, not a way to keep a page out of Google. Use a noindex directive or password protection for pages which must stay out of search results, for example a staging site. Step 3 is an XML sitemap. In 2026, one sitemap file holds at most 50,000 URLs or 50 MB uncompressed, and Google ignores the priority and changefreq values. The lastmod date only counts when the date is consistently and verifiably accurate. Finally, return a real 404 or 410 status for removed pages and a 301 redirect for moved pages, with Next.js or any other framework.
How do you handle duplicate URLs, canonicals and hreflang?
A canonical URL is the preferred version of a page when several URLs show the same content. A redirect and a rel="canonical" annotation are both strong signals, according to Google, while inclusion in a sitemap is a weak signal. Step 1 is one URL format, enforced everywhere: with or without trailing slash, with or without www, and always HTTPS. Do not use robots.txt for canonicalisation, because Google may still index blocked URLs without the content. Step 2 is hreflang for multilingual sites. Google accepts hreflang in HTML link tags, HTTP headers or the XML sitemap, and treats the three methods as equivalent. Each language version must list itself and every other version. When two pages do not point to each other, Google ignores the tags. Additionally, add an x-default for visitors whose language matches no version. For example, a site with 20 pages in three languages gives every page four hreflang links. The virtuolab.dev website, built with Next.js, handles hreflang the same way in the sitemap.
Does JavaScript rendering hurt SEO?
JavaScript rendering is the step in which Google runs the scripts of a page in a headless Chromium browser before indexing the result. Google renders pages once resources allow, so content which only appears after client-side scripts can reach the index later than server-rendered HTML. In 2026, server-side rendering or pre-rendering is still a great idea, according to Google, because the site becomes faster for users and crawlers. Not all bots can run JavaScript either. For example, link previews in chat apps usually read only the raw HTML. A site built with Next.js avoids the problem through server components and static generation, which send complete HTML on the first request. A single-page app in plain React needs pre-rendering or a move to a framework. Step 1 of every check is the URL Inspection tool in Google Search Console. The tool shows how Googlebot sees the page and which resources failed to load.
Which Core Web Vitals thresholds should you hit in 2026?
Core Web Vitals are three metrics from Google which measure the loading speed, responsiveness and visual stability of real page visits. According to Google, a good page reaches a Largest Contentful Paint (LCP) within 2.5 seconds, an Interaction to Next Paint (INP) of 200 milliseconds or less, and a Cumulative Layout Shift (CLS) of 0.1 or less. Google measures the thresholds at the 75th percentile of page loads, split by mobile and desktop. A site passes only when at least 75% of real visits meet the target. The HTTP Archive Web Almanac 2025 shows a good LCP on 62% of mobile pages, against 74% on desktop. In practice, LCP problems often come from a large hero image, INP problems from heavy JavaScript and CLS problems from images without dimensions. Step 1 is field data in Google Search Console or PageSpeed Insights. Step 2 is debugging with Lighthouse.
Which structured data is still worth adding in 2026?
Structured data is machine-readable markup, usually JSON-LD with schema.org types, which describes what a page is about. For most business sites, Organization, BreadcrumbList and Article remain the useful core, with Product for webshops. The landscape changed in 2026. Since May 2026, FAQ rich results no longer appear in Google Search, according to Google. Google removed the FAQ documentation in June 2026. FAQ markup does no harm, but nobody should expect extra space in the results anymore. Additionally, structured data is not required for generative AI search, and no special schema.org markup exists for AI features, according to Google documentation. Google still recommends structured data as part of an overall SEO strategy. In practice, keep JSON-LD consistent with the visible text. First, generate the markup from the same data source as the page. Second, validate every template with the Rich Results Test before each release.
How does technical SEO affect visibility in AI search?
Visibility in AI search means being shown or cited in answers, for example through Google AI Overviews and AI Mode. Step 1 is technical. A page must be indexed and eligible for a snippet in Google Search before generative AI features can show the page, according to Google. Google calls optimising for generative AI search simply SEO, and says no new machine-readable files or special markup are needed. A Pew Research study from 2025 found an AI summary on 18% of the Google searches measured. Users clicked a result link in 8% of visits with a summary, against 15% without one. Being cited inside the answer itself therefore becomes increasingly valuable. Additionally, ChatGPT and other assistants are separate systems, so clean server-rendered HTML is the safest common denominator. Generative engine optimisation (GEO) adds clear answers, consistent company facts and named sources. VirtuoLab helps companies get mentioned faster and more often in Google and in AI assistants such as ChatGPT.
Frequently asked questions
Does a small website still need an XML sitemap?
An XML sitemap is not strictly required for a small website with good internal links, but the sitemap remains cheap insurance. Google uses sitemap inclusion as a weak signal for the canonical URL, according to Google. A sitemap also speeds up discovery of new pages, for example a fresh blog article without many internal links yet. Additionally, keep the file honest. Google ignores priority and changefreq, and uses lastmod only when the date is consistently and verifiably accurate. A lastmod which changes on every build therefore carries no value. A sitemap.ts file in a site built with Next.js generates the sitemap from the same data as the pages, so new URLs appear automatically. Step 1 is to submit the sitemap once in Google Search Console. Step 2 is a monthly check of the indexing report. Even for a site with 20 pages, the small setup effort keeps paying off in 2026.
Is Next.js good for technical SEO?
Next.js is a React framework which renders HTML on the server or at build time. The framework is a strong starting point for technical SEO. Server-rendered HTML also matches the advice from Google for JavaScript sites. The framework covers three crawl basics out of the box. Step 1 is the metadata API, which sets titles, descriptions, canonical URLs and hreflang alternates per page. Step 2 is a pair of special sitemap and robots files, which generate both crawl files from code. Step 3 is the image component, which requires dimensions and protects the CLS score. The framework does not fix weak content, broken internal links or slow third-party scripts, however. Tag managers, chat widgets and video embeds often cost more INP than the framework itself. React, the library underneath, was used by 44.7% of respondents in the Stack Overflow Developer Survey 2025. In practice, VirtuoLab builds websites and web apps on the same framework with TypeScript and PostgreSQL, including the VirtuoLab website itself.
Should you add an llms.txt file to your website?
An llms.txt file is a plain text summary of a website for language models, placed at the root of the domain. Such files are not needed for Google Search, according to Google, and Google Search ignores the files. The file does not help or hurt rankings in Google. Other services and AI tools may still read the file, so the summary can make sense as a small extra. For example, the VirtuoLab website, built with Next.js, publishes one with a company summary in Dutch, English and German. Treat the file as optional, never as a replacement for the basics. A page which is not indexed, renders without content or loads slowly will not become visible through a text file. Step 1 in 2026 is therefore crawling, rendering and Core Web Vitals. Step 2 is structured data. Finally, add an llms.txt if time allows.