HTML Meta Tags Cheatsheet - Meta Tags SEO & Configuration Reference

Yes, both, because they serve different consumers. title is what search engines and the browser tab read, while og:title is what social platform crawlers pick up when building a share card — those crawlers only look at Open Graph and Twitter Card properties. Writing both does not count as duplicate content since they render in different places. The real care point is that they should not contradict each other: an og:title far from the visible page title makes the shared card look unlike the landing page.

Each is evaluated independently, and they are not redundant: the robots meta applies to every crawler that follows the generic convention, while googlebot targets only Google. Writing googlebot with noindex alone is therefore enough to keep Google out while other engines still index the page, which is exactly the case the distinction exists for. If the two carry opposite instructions, behaviour diverges per crawler, and that split is among the hardest things to diagnose when indexing misbehaves — in practice, either make them agree or drop one entirely.

Point it at the parameterless canonical address. URLs carrying UTM tags, sort order or filter parameters can generate a large number of equivalent pages, and canonical exists to funnel all of them into one authoritative address so ranking signals are not split. hreflang operates at a different layer: a multilingual site declares each language version with alternate hreflang and marks the fallback with x-default. Canonical answers "same content, several addresses", hreflang answers "different languages, different addresses" — they coexist independently.

Only to a limited degree, since caching is primarily governed by response headers. A meta http-equiv Cache-Control is honoured only in some cases, and only when it appears early in the document — placed later it tends to be ignored. Its real value is in static exports or local file previews where no genuine response header exists, which makes it handy for temporary debugging and forcing a fresh fetch. A production site should emit Cache-Control as a response header and treat the meta variant as a supplement rather than the main mechanism.

Yes, but the light value needs a media attribute so it only applies to the matching preference, with the dark value left unconditioned as the fallback. Dark-mode users are keyed off their system setting, so two rules separated by prefers-color-scheme cover both cases within one group. The toolbar colour is purely decorative and getting it wrong breaks nothing functionally, but a bright value on a dark background is glaring — so this tag should switch with your dual-theme work instead of being hard-coded globally.