Google's web.dev guidance recommends keeping TTFB under 800 milliseconds, since it precedes every other loading milestone — a page cannot start rendering before its first byte arrives. Most well-configured sites with CDN caching land in the 100-500ms range.

Why do AI crawlers care about TTFB?

AI fetchers operate under strict time and resource budgets, and many fetch pages live during answer generation, where a user is waiting. A slow-responding origin risks timeouts, partial fetches, and deprioritized recrawling — meaning slow pages can silently drop out of the retrieval pool.

What Is Time to First Byte (TTFB)?

Last updated: July 5, 2026 · Menra Team

Time to First Byte (TTFB) is the interval between a client requesting a page and receiving the first byte of the response. It bundles DNS resolution, connection and TLS setup, server processing, and network transit. Google's web.dev guidance treats 800 milliseconds as the threshold for good TTFB, because it gates every downstream metric — Largest Contentful Paint's 2.5-second budget starts spending itself before your server has said a word.

Why is TTFB a GEO metric now?

AI crawlers are less patient than human visitors. Bulk crawlers like GPTBot and ClaudeBot work through millions of URLs under crawl-budget economics: origins that respond slowly get fetched less often and less deeply. On-demand fetchers — ChatGPT-User or PerplexityBot retrieving a page mid-answer — operate with a user actively waiting, so their timeout tolerance is tight; an origin that takes seconds to respond can miss the answer entirely. In both modes, TTFB failures are invisible: no error appears anywhere on your side except a log line, while the engine quietly cites a faster competitor.

What actually moves TTFB?

CDN caching. Serving HTML from edge cache turns a 900ms origin round trip into double-digit milliseconds — the single largest lever.
Server processing. Uncached database queries, slow middleware, and cold serverless starts dominate origin time.
Compression and protocols. HTTP/2 or HTTP/3 plus Brotli reduce connection overhead.
Geographic distance. Origin-only hosting means intercontinental requests pay full latency; edge distribution fixes it.
Measure like a bot: test uncached, from multiple regions, with crawler user agents — your warmed browser cache lies to you.

Example

A docs site on a single-region origin shows 1.4s TTFB from other continents. After fronting it with a CDN that caches HTML, TTFB drops under 200ms globally — and server logs show AI crawler hit depth increasing over the following weeks.

Related terms

See crawler timeout, Core Web Vitals, and edge SEO; thresholds are documented at web.dev. Performance fixes rank early in the GEO optimization sequence.

Frequently asked questions

What is a good TTFB?: Google's web.dev guidance recommends keeping TTFB under 800 milliseconds, since it precedes every other loading milestone — a page cannot start rendering before its first byte arrives. Most well-configured sites with CDN caching land in the 100-500ms range.
Why do AI crawlers care about TTFB?: AI fetchers operate under strict time and resource budgets, and many fetch pages live during answer generation, where a user is waiting. A slow-responding origin risks timeouts, partial fetches, and deprioritized recrawling — meaning slow pages can silently drop out of the retrieval pool.

Keep exploring

See how AI engines talk about your brand — track mentions across ChatGPT, Perplexity, Claude, Gemini and 5 more. Start with Menra