AnalysisNew

    Lovable Renders on the Server Now. What That Fixes, and What It Doesn't.

    Server rendering solves one problem precisely: a crawler that fetches your page gets your words. It does not, on its own, hand you per-route titles, canonicals, a sitemap with real dates, or cards a crawler can follow.

    Not sure what is wrong? Read why Google is not indexing your Lovable site and how to fix it.

    On 28 July 2026, Lovable's changelog added a line that changed the first question we ask on an SEO call. Projects built on the older React + Vite stack can now upgrade in place to TanStack Start, which, in Lovable's own words, “renders your pages on the server so search engines and social link previews can read your content.”

    That is a real fix for a real problem, and it removes the single most common reason founders used to call us. It is also narrower than it sounds. Server rendering answers one of the questions a crawler asks about your page. There are about six.

    What the upgrade actually fixes

    A client-rendered app returns an HTML shell and fills it in once JavaScript has run. Server rendering means the words are already in the response. Lovable's own framing is the useful one: the point of the change is that search engines and social link previews can read your content. If your pages were arriving empty, this fixes that, and it fixes it at the root rather than by bolting a prerenderer onto the side.

    It is worth being precise about what that buys you, because the gap between readable and indexed well is where most of the work we get called in for actually lives.

    The upgrade is opt-in, so the first question is about your project

    Note Lovable's phrasing: older projects “can now upgrade in place”. Nothing moved on its own. A project created before the switch is still on React + Vite until somebody presses the button, which means “does Lovable render on the server?” is the wrong question. The right one is “does mine?” If you are not sure, our seven-point indexability check walks through how to tell.

    What a crawler needs, and where server rendering stops

    What a crawler needsSupplied by the server-rendering upgradeStill yours to build
    Page content present in the HTML responseYes — this is the changeNothing further, once your project is on the new stack
    A distinct title and description per routeNoPer-route metadata, not one set repeated sitewide
    A self-referencing canonical on every pageNoOne canonical per URL, pointing at itself
    A sitemap whose lastmod moves when content changesNoGenerated at build, dated from the content
    Structured data (Article, BreadcrumbList)NoAdded once per template
    Internal links a crawler can followNoReal anchor elements, not clickable divs

    The row that catches people

    The last row is the one we find most often, and it survives the upgrade untouched. A card that looks like a link but is a div with a click handler is not a link. A crawler reading your HTML sees a box of text with nowhere to go, so the page it should have discovered stays undiscovered. Server rendering does not make markup good. It delivers it sooner.

    The same goes for the rest of that column. Rendering on the server does not decide what your title tag says on a given route, does not declare which URL is canonical when two paths serve the same content, and does not write a sitemap whose lastmod moves when you edit something. Those stay build-time decisions somebody has to make once per template.

    So do you still need to migrate?

    Sometimes the honest answer is no, and we would rather say that on a call than after an invoice. If client-side rendering was the only thing standing between you and being indexed, and the rest of the list above is already handled, the in-place upgrade may be the whole fix.

    The reasons that survive it are the ones that were never about rendering.

    1. Owning the code. Lovable added a codebase download on 14 August 2026 — "a one-time copy of your project as a .zip file, no repository connection needed". That is a copy. It is not a repository, a review process and a deploy pipeline you control.
    2. Owning the patch cycle. Next.js moved to a formal security release process in July 2026, publishing patches on a regular schedule with advance notice, and releases followed in July, August and September. Running Next.js means subscribing to that cadence, and somebody has to be the subscriber.
    3. Everything your database is about to ask of you, which is identical on either stack.

    The deadline that does not care which stack you are on

    On 28 April 2026 Supabase published a breaking change: new tables in the public schema are no longer exposed to the Data and GraphQL API by default. It became the default for new projects on 30 May 2026, and it is enforced on all projects on 30 October 2026. That is about five weeks from today.

    If anything you shipped relies on a table being reachable simply because it exists in the public schema, that is the check to run this month rather than next. It is not a big job. It is a job with a date on it, and the date does not move because you upgraded your rendering. The same is true of what is queued behind it: supabase-js requires TypeScript 5.0 from 31 January 2027, and extension version pinning has been ignored since 5 August 2026.

    Where we land

    Our own numbers, for the record: 80 jobs on Upwork, 100% Job Success and a 4.9 rating from 30 reviews. The Strategy Audit is $499, one-time, and it credits toward any build — including the builds it tells you not to do. If the audit concludes that the in-place upgrade plus a morning of metadata work is your whole SEO problem, that is what the written roadmap will say. If you do need the rest, Founding Client is $2,999 with a 14-day build guarantee and a full refund if we miss the deadline, and Publishable MVP is $4,999. Both include production deployment, monitoring, documentation and handoff, and 90-day support until your first paying customer, and the revenue guarantee is first paying customers within 90 days of launch, or full refund.

    Server rendering closed a gap that used to cost founders months. The rest of the list is still a list.

    If you are not sure which side of it you are on, book your free 20-minute intro call.

    Sources

    Related posts