Updated 2026-07-13 · 11 min read
A Lovable project is not search-optimized because it has a title field and looks good in preview. You need to inspect the deployed output, define which routes deserve indexing and give every page its own metadata, internal links and structured data. Use this guide before submitting the sitemap.
Visit the production URL and view source. Search for the page’s H1 and a paragraph of body copy. If they are absent until JavaScript runs, Google may eventually render them, but discovery and AI-crawler visibility are less reliable. For a content-heavy project, consider exporting to a deployment that supports static generation or prerendering, or generate crawlable snapshots for public routes.
Create a list of public pages, their intent and canonical URL. Keep account, admin, checkout, API and internal search states out of the index. For dynamic pages, define a stable slug and require enough unique content to justify the page. Programmatic pages with only a swapped noun are thin content regardless of the builder.
For every public route, set a unique title under ~60 characters, a useful meta description, canonical URL, Open Graph title/description/image/URL and one H1 matching the page intent. Derive dynamic values from the route data. Never reuse the homepage metadata. Add noindex to account, admin, checkout and empty search routes.
Primary navigation, related content and breadcrumbs should use anchor elements with real href values. Do not rely exclusively on click handlers or hidden menus. Every important page should receive internal links from a hub and relevant sibling pages.
Generate the sitemap from the canonical route inventory and expose it at /sitemap.xml with an XML content type. Reference it in robots.txt. Submit the full URL to Search Console and inspect several representative routes individually. “Submitted” does not mean “indexed”; investigate duplicate, discovered-not-indexed and crawled-not-indexed states separately.
Discovered — currently not indexed often means Google knows the URL but does not yet consider crawling it worthwhile; strengthen internal links and consolidate thin routes. Crawled — currently not indexed means the content was seen but not selected; check duplication, usefulness and canonical signals. Duplicate without user-selected canonical points to conflicting URL variants or missing canonicals. Inspect the rendered production page, not only the editor preview.
After a fix, request indexing for a small representative sample. If those pages improve, roll the same template correction across the set and let the sitemap handle discovery.
Add organization and website entities globally, then page-specific types: Article for real articles, BreadcrumbList for visible breadcrumbs, SoftwareApplication only when the page genuinely describes software and its offer, FAQPage only when the questions and answers are visible. Validate the rendered JSON-LD and keep entity IDs consistent.
Compress generated hero assets, set width/height, limit webfont files and remove decorative libraries that add no user value. Test on a real phone and slow network. Lovable’s preview speed on a desktop is not field performance.
Audit this deployed project for SEO without changing visual design. Inventory all public routes. For each, report rendering method, unique title/description/canonical/H1, crawlable internal links, indexability and JSON-LD. Verify sitemap and robots alignment, image alt/dimensions, semantic landmarks and Core Web Vitals risks. First return a prioritized report; do not edit. After approval, fix one category at a time and verify the production HTML after each change.
Yes. Verify crawlable content, unique page metadata, internal architecture, performance and useful content exactly as you would for any React-based site.
Ask the agent to generate sitemap.xml from your canonical public route list, then verify the production URL returns valid XML and submit it in Search Console.
Common causes include client-only content, duplicate metadata, weak/thin pages, missing internal links, canonical mistakes or simply insufficient time and authority. Search Console’s URL inspection narrows it down.