Google's crawler skips JSON. Its indexer still reads it

Google's crawler does not read JSON. Gary Illyes, a Google engineer, said on Bluesky this morning that Googlebot only downloads JSON files, and that parsing happens later, in indexing.

"Google's crawlers don't parse JSON, they just download things. Parsing anything is done by, for example, indexing in case of Search," he wrote.

That distinction matters for anyone using JSON-LD to shape how a brand shows up in AI Overviews. The crawler is not the step that reads your schema markup. The indexer is downstream, and it still processes it. Nothing here says structured data goes unread by Google. It says the crawler specifically is not the one reading it.

Illyes also confirmed the same applies to Merchant Center, when asked directly. Google's crawlers share fetching infrastructure across services, so the download-only behavior is not unique to Search. He said he was "very certain" Merchant Center uses the same parser, though he added he has never worked with that team directly, so treat that specific claim as slightly softer than the Search confirmation.

What this doesn't tell us

The clarification describes Google's own pipeline. It says nothing about how other AI systems that crawl the web, ChatGPT's crawler or Perplexity's, for instance, handle JSON-LD at the point they fetch a page. Some of those systems may parse structured data during the fetch itself rather than in a separate step. Until that's confirmed, the practical advice does not change: keep structured metadata complete, because a company selling to AI-driven search does not control which stage of which company's pipeline actually reads it.