fix(webapp): use blocking fallback for ISR pages to fix SEO

[nimrodkra]

Summary

• Switch fallback: true → fallback: 'blocking' on posts, tags, and sources pages
• With fallback: true, crawlers receive an empty skeleton HTML (~2.5KB) with generic title, no canonical, no OG tags, no H1, and no content
• With fallback: 'blocking', Next.js waits for getStaticProps to complete before responding, ensuring crawlers always get the fully rendered page (~12KB+) with all SEO metadata

Context

Ahrefs site audit (Mar 5) found ~18,753 pages flagged across multiple SEO categories — all caused by the same root issue:

Issue	Count
Orphan pages (no internal links)	19,503
Missing H1 tag	18,958
Low word count	18,796
Duplicate pages without canonical	18,753
Incomplete OG tags	18,753
No outgoing links	18,754

All are post/tag/source pages from sitemaps that Ahrefs crawled during ISR fallback, receiving skeleton HTML instead of rendered content.

Test plan

• Verify posts, tags, and sources pages still load correctly
• Check that first-visit to an uncached post/tag/source serves full HTML (not skeleton)
• Confirm meta tags, canonical, OG tags, and H1 are present in initial HTML response
• Monitor TTFB — first requests to uncached pages will be slightly slower (blocking instead of immediate skeleton), but subsequent requests are unchanged

🤖 Generated with Claude Code

Preview domain

https://fix-isr-fallback-blocking-seo.preview.app.daily.dev

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
daily-webapp	Ready	Preview	Mar 8, 2026 10:12am

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
storybook	Ignored		Mar 8, 2026 10:12am

[nimrodkra]

@claude please review this pr

[claude]

Claude finished @nimrodkra's task in 1m 30s —— View job

---

PR Review

• Gather context and understand the request
• Read changed files
• Analyze the changes
• Post review feedback

---

Summary

The fix is correct and well-motivated. Switching from fallback: true to fallback: 'blocking' is the right approach to fix the SEO issue — crawlers will now always receive a fully-rendered page with all metadata instead of a skeleton HTML.

---

Observations

✅ Core change is correct

All three getStaticPaths functions are correctly updated. With fallback: 'blocking', Next.js will SSR the page on first request and cache it, so any crawler hitting an uncached URL gets the full HTML immediately.

⚠️ Dead code: isFallback checks are now unreachable

With fallback: 'blocking', router.isFallback is always false — Next.js never puts the page in the fallback state. The existing isFallback checks in all three components are now dead code:

posts/[id]/index.tsx line 216:

if (isLoading || isFallback || privateSourceJoin.isActive) {
  // isFallback is always false now

sources/[source].tsx lines 162–167:

if (!isFallback && !source) {      // simplifies to: if (!source)
  return <Custom404 />;
}
if (isFallback || !source) {       // isFallback branch is dead
  return <></>;
}

tags/[tag].tsx lines 234–246:

if (isFallback) {                  // always false now — entire block is dead
  return (
    <FeedPageLayoutComponent>
      <PageInfoHeader ...>...</PageInfoHeader>
    </FeedPageLayoutComponent>
  );
}

These dead branches won't cause bugs, but they're misleading and should be cleaned up as a follow-up (or in this PR). Fix this →

ℹ️ Pre-existing: Sources page returns soft 404

sources/[source].tsx getStaticProps (line 311) returns { props: { source: null }, revalidate: 60 } for not-found/forbidden sources, instead of { notFound: true }. This means the response is HTTP 200 with a Custom404 component — a "soft 404" that search engines may not treat as a true 404. The posts page already uses notFound: true correctly at line 348–351. Not introduced by this PR, but worth a follow-up for SEO consistency. Fix this →

ℹ️ TTFB impact acknowledged

First-time requests to uncached pages will block until getStaticProps completes. Given the parallel API fetches already in place (posts page fetches data and comments concurrently), this should be acceptable. Worth monitoring after deploy.

---

Verdict

LGTM — the fix is minimal, correct, and directly addresses a real SEO regression. The isFallback cleanup is a nice-to-have but not a blocker.

[nimrodkra]

@claude let's run a quick 2nd round of review

[claude]

Claude finished @nimrodkra's task in 1m 20s —— View job

---

Second Round Review

• Read changed files
• Verify isFallback cleanup
• Verify sources soft 404 fix
• Post review feedback

---

All issues flagged in the first review have been addressed. Here's the second-pass breakdown:

✅ sources/[source].tsx — fully cleaned up

• useRouter import removed (no longer needed) ✅
• isFallback removed from the component entirely ✅
• Dead if (isFallback || !source) return <></>; block removed ✅
• Soft 404 fixed: getStaticProps now returns { notFound: true, revalidate: 60 } instead of { props: { source: null } } for not-found/forbidden sources ✅

✅ tags/[tag].tsx — fully cleaned up

• isFallback removed from useRouter() destructure ✅
• Entire fallback skeleton UI block removed ✅

⚠️ posts/[id]/index.tsx — cleanup is partial but intentional

The isFallback variable is now const isFallback = false; (hardcoded) rather than removed:

// line 167
const isFallback = false;

// line 253
<Content ... isFallback={isFallback} ... />

This is correct because isFallback prop still flows through the Content → PostContent → BasePostContent chain, and BasePostContent uses it as a guard:

// BasePostContent.tsx:39
if (!id && !isFallback) {
  return <Custom404 />;
}

With fallback: 'blocking', post will always be hydrated server-side so id will always exist, making isFallback permanently false and harmless. The hardcoded false is an honest representation of the new runtime behavior. Full cleanup of the isFallback prop from the shared component tree (PostContent, BasePostContent, SquadPostContent, etc.) would be a broader refactor — fine to defer.

---

Verdict

LGTM — the second commit cleanly addresses all first-round feedback. The fix is minimal, correct, and all three pages are in good shape.