google, any year you like

A time machine for the web. Pick a year from 1998 to 2013, get that era's Google — the real layouts, logos, tabs, sidebars and pagination — and search. Every result opens the page as it looked in that year, served by the Internet Archive's Wayback Machine, so you keep browsing the old web from there.

Google's actual rankings from past years aren't retrievable — nobody archived live search results. This is an approximation: Google Custom Search restricted to pages dated on or before your chosen year (when the daily free quota allows), with the Internet Archive's site index — filtered to sites that were being captured in that year, ranked by capture count — as the always-on fallback. Each results page says which source it used.

Nothing is logged — no queries, no IPs, no analytics. A per-visitor daily budget (counted, not logged) keeps the free API quotas alive. Repeat searches are served from an edge cache.